Therapeutic hypothermia and controlled normothermia in the ... · Therapeutic hypothermia and...

Review Article

Therapeutic hypothermia and controlled normothermia in theintensive care unit: Practical considerations, side effects, andcooling methods*

Kees H. Polderman, MD, PhD; Ingeborg Herold, MD

I nduced (therapeutic) hypother-mia, defined here as an inten-tional reduction of a patients’ coretemperature to 32°C–35°C (Table

1), is being used with increasing fre-quency as a method to prevent or miti-gate various types of neurologic injury(1). In recent years there has been a sig-nificant increase in our understanding ofthe cascade of destructive processes thatunfold in the injured brain in the min-utes to hours after an episode of ischemiaor trauma. These processes, which havebeen collectively termed postresuscita-

tion disease and reperfusion injury in thecase of postanoxic injury and as “second-ary brain injury” in the case of traumaticinjury (TBI), can continue for hours toseveral days after the initial injury, andcan be retriggered by new episodes ofischemia. The key point is that all ofthese processes are temperature depen-dent; they are all stimulated by fever, andcan be blocked or mitigated by mild tomoderate hypothermia. The wide-rang-ing effect of hypothermia on all of thesemechanisms may explain why therapeu-tic hypothermia has proved to be clini-cally effective, whereas studies with phar-macologic agents that affect just one ofthe destructive processes have beenmuch less successful (1).

The most recent guidelines from theAmerican Heart Association and the Euro-pean Resuscitation Council recommendthe use of hypothermia for selected patientswho remain comatose following a wit-nessed cardiac arrest (2). Under certain

conditions, hypothermia is also used ther-apeutically in the treatment of severe trau-matic brain injury, stroke, hepatic failure,spinal cord injury, myocardial infarction,and numerous others. The evidence forthese and other potential indications hasbeen discussed in various reviews (1, 3, 4).

Another important development inthe field of neurocritical care is the in-creasing awareness that development offever can adversely affect outcome inneurocritical patients. Various studieshave shown that fever is independentlyassociated with adverse outcome in pa-tients with neurologic injury (includingpostanoxic injury following cardiac arrest),regardless of the cause of fever (1, 5–9).

All this means that the issue of tem-perature control and temperature manip-ulation is gaining importance in neuro-critical care. However, inducing hypothermiaand/or maintaining normothermia in-duces a large number of changes thatcan pose significant risks, and manipu-

*See also p. 1172.From the Department of Intensive Care, University

Medical Center Utrecht, The Netherlands.The authors have not disclosed any potential con-

flicts of interest.For information regarding this article, E-mail:

[email protected] or [email protected] © 2009 by the Society of Critical Care

Medicine and Lippincott Williams & Wilkins

DOI: 10.1097/CCM.0b013e3181962ad5

Background: Hypothermia is being used with increasing fre-quency to prevent or mitigate various types of neurologic injury.In addition, symptomatic fever control is becoming an increas-ingly accepted goal of therapy in patients with neurocriticalillness. However, effectively controlling fever and inducing hypo-thermia poses special challenges to the intensive care unit teamand others involved in the care of critically ill patients.

Objective: To discuss practical aspects and pitfalls of thera-peutic temperature management in critically ill patients, and toreview the currently available cooling methods.

Design: Review article.Interventions: None.Main Results: Cooling can be divided into three distinct phases:

induction, maintenance, and rewarming. Each has its own risksand management problems. A number of cooling devices thathave reached the market in recent years enable reliable mainte-nance and slow and controlled rewarming. In the induction phase,rapid cooling rates can be achieved by combining cold fluidinfusion (1500–3000 mL 4°C saline or Ringer’s lactate) with aninvasive or surface cooling device. Rapid induction decreases therisks and consequences of short-term side effects, such as shiv-

ering and metabolic disorders. Cardiovascular effects includebradycardia and a rise in blood pressure. Hypothermia’s effect onmyocardial contractility is variable (depending on heart rate andfilling pressure); in most patients myocardial contractility willincrease, although mild diastolic dysfunction can develop in somepatients. A risk of clinically significant arrhythmias occurs only ifcore temperature decreases below 30°C. The most importantlong-term side effects of hypothermia are infections (usually ofthe respiratory tract or wounds) and bedsores.

Conclusions: Temperature management and hypothermia in-duction are gaining importance in critical care medicine. Intensivecare unit physicians, critical care nurses, and others (emergencyphysicians, neurologists, and cardiologists) should be familiarwith the physiologic effects, current indications, techniques, com-plications and practical issues of temperature management, andinduced hypothermia. In experienced hands the technique is safeand highly effective. (Crit Care Med 2009; 37:1101–1120)

KEY WORDS: hypothermia; therapeutic; definitions; fever control;normothermia; side effects; neurologic injury; cardiac arrest;traumatic brain injury

1101Crit Care Med 2009 Vol. 37, No. 3

lating body temperature in critically illpatients in a safe way can present con-siderable challenges to intensive carephysicians.

This review will address the physio-logic changes and potential side effectsassociated with hypothermia. Currentlyavailable methods for inducing hypother-mia will also be discussed, and practicalrecommendations on how to deal withpotentially harmful effects, as well as pre-ventive measures, will be provided to helpguide clinicians through this sometimescomplex treatment.

One of the issues leading to misunder-standings in the field of therapeutic tem-perature management has been confu-sion over terminology. Terms such astherapeutic hypothermia and inducedhypothermia have been used with differ-ent meanings, and expressions such asmild/moderate/severe hypothermia havebeen used to describe widely differenttemperature ranges. In addition, re-search findings from the field of acci-dental hypothermia have often (andsometimes inappropriately) been di-rectly translated to the field of thera-peutic hypothermia, without takinginto account some major differencesbetween these two clinical situations.For example, accidental hypothermia inthe perioperative setting leads to anincrease in heart rate, whereas con-trolled therapeutic hypothermia leadsto a decrease in heart rate. These issuesare discussed further below; a list of

proposed definitions is provided inTable 1.

Literature Search Strategies

An electronic search of MEDLINE(OVID), EMBASE, Current Contents, andthe Cochrane library was performed. Thesearch terms used included “hypother-mia“ or “cooling” in various combina-tions with “methods,” “devices,” “induc-tion,” “arrhythmias,” “hemodynamic,”“bleeding,” “side effects,” “coagulopathy,”“infection,” “immune suppression,” andvarious others. One of the authors has apersonal archive with �1000 papers onthe subject of hypothermia and tempera-ture manipulation; data from this archiveand from two reference books that havebeen published on the subject of inducedhypothermia were also used (10, 11). Ahand search of key journals was per-formed for studies published before 1966.The search was not restricted by languageor type of publication. Studies from thefield of perioperative hypothermia and ac-cidental hypothermia were included inthe search. Where no published studieswere available the authors drew on theirexperience in �1000 patients. Where rec-ommendations can be made based onpublished studies this is indicated by theappropriate references.

Induction of Hypothermia

Attempts to induce hypothermia willlead to the activation of counter-regula-

tory mechanisms to decrease heat loss.Under normal circumstances this will beaccomplished by increasing the sympa-thetic tone and through vasoconstrictionof vessels in the skin (12). This responsecomplicates attempts to induce therapeu-tic hypothermia by surface cooling (seebelow). Under normal circumstances va-soconstriction begins at a core tempera-ture of around 36.5°C (13); the reductionin heat loss resulting from cutaneous va-soconstriction is �25% (14). In addition,heat production will be increasedthrough shivering, with the shiveringthreshold being �1°C below the vasocon-striction threshold (so at �35.5°C) (13).Shivering has been linked to an increasedrisk of morbid cardiac events and adverseoutcome if it occurs in the postoperativephase, where the patient who has becomehypothermic and shivers will have a highrate of metabolism, increased oxygenconsumption, excess work of breathing,higher heart rate, and a general stress-like response (15–18). In awake patients,this type of full counter-regulatory re-sponse can increase oxygen consumptionby between 40% and 100% (19–24). Thisis an undesirable effect particularly inpatients with neurologic and/or posthy-poxic injury; indeed accidental peri-operative hypothermia and the resultingstress response have been linked to anincreased risk of morbid cardiac events,particularly in older patients with heartdisease (15–17).

It should be noted that these adverseeffects are linked to the hemodynamicand respiratory responses rather than tothe shivering, per se (15). When an awakepatient develops perioperative hypother-mia, the average heart rate increases sig-nificantly (15–18); in contrast, inducinghypothermia intentionally in sedated pa-tients has the opposite effect, with signif-icant decreases in heart rate (12, dis-cussed more extensively later). Shiveringwill increase oxygen consumption, but asthe patient is on mechanical ventilation,there will be no increase in the work ofbreathing. The main problem of shiver-ing during the induction phase is that itgenerates significant amounts of heat;sustained shivering can double the met-abolic rate (12, 19), thereby significantlydecreasing cooling rates. For this andother reasons, shivering should be ag-gressively treated especially in the induc-tion phase of cooling (see later).

Shivering can be counteracted by ad-ministration of sedatives, anesthetics,opiates, magnesium, muscle paralyzers,

Table 1. Proposed terms and definitions surrounding therapeutic hypothermia

Therapeutic temperature management definitionsHypothermia Core temperature �36.0°C regardless of the causeInduced hypothermia An intentional reduction of a patients’ core temperature

below 36.0°CTherapeutic hypothermia Controlled induced hypothermia: i.e., induced

hypothermia with the potentially deleterious effects,such as shivering, being controlled or suppressed

Controlled normothermia/therapeuticnormothermia

Bringing down core temperature in a patient with fever,and maintaining temperature within a range of36.0°C–37.5°C, with the potentially deleterious effects,such as shivering, being controlled or suppressed

Temperature range definitionsMild therapeutic hypothermia An intentional and controlled reduction of a patients’

core temperature to 34.0°C–35.9°CModerate therapeutic hypothermia An intentional and controlled reduction of a patients’

core temperature to 32.0°C–33.9°CModerate/deep therapeutic hypothermia An intentional and controlled reduction of a patients’

core temperature to 30.0°C–31.9°CDeep therapeutic hypothermia An intentional and controlled reduction of a patients’

core temperature to �30.0°CMild hyperthermia Core temperature 37.5°C–38.0°CModerate hyperthermia Core temperature 38.1°C–38.5°CModerate/severe hyperthermia Core temperature 38.6°C–38.9°CSevere hyperthermia Core temperature �39.0°C

1102 Crit Care Med 2009 Vol. 37, No. 3

and various other drugs (discussed laterand in Table 4). Skin counterwarming ofthe hands, feet, and face can also be usedto reduce shivering (60, 99). In our expe-rience, shivering can be managed with-out the use of paralytic agents in mostpatients in the setting of intensive care.

There are several reasons to reducethe use of paralytic agents during cool-ing. First, although muscular shiveringactivity will be blunted by paralysis, therewill be no effect at the central level; inother words, the brain’s attempts to gen-erate a shivering response will not cease.Second, administration of paralyticagents may mask seizure activity. Sei-zures may occur in a substantial propor-tion of patients after postanoxic injury(25) or TBI (26), and unless patients areroutinely monitored for seizure activityusing continuous electroencephalo-graphic monitoring paralysis can causediagnostic problems. Third, using seda-tives and analgesics to combat shiveringhas the advantage of inducing vasodila-tion, thereby increasing transfer of heatfrom the core to the periphery and fur-ther (1). This phenomenon occurs notonly with volatile anesthetics but alsowith intravenous anesthetics and analge-sics (32–38). Fourth, it is well recognizedthat prolonged paralysis significantly in-creases the risk of critical illness polyneu-romyopathy (42); thus, in general, itmakes sense to avoid prolonged adminis-tration of these agents if possible. Finallyand perhaps most importantly, paralysismay mask insufficient sedation. Animalexperiments have shown that the protec-tive effects of hypothermia can be par-tially or even completely lost when ani-mals are not sedated during hypothermiatreatment. In one example, Thoresen et al(27) exposed newborn piglets to a periodof global anoxia, after which some ani-mals were treated with hypothermia. Nosedation or analgesia was given to eitherhypothermic animals or controls. No pro-tective effects on neurologic outcomewere observed in this study (27). How-ever, when the same research group per-formed the same experiment in the sameanimal model, with the same type andduration of injury but with sedation andanalgesia used in both groups, there wasa large improvement in neurologic out-come associated with hypothermia treat-ment (28). The loss of protective effects in thefirst study (27) was attributed by the authorsto an aggravated stress response, which wasprevented by appropriate sedation and anal-gesia in the second study (28).

Most clinical studies that have appliedhypothermia successfully have used deepsedation and analgesia in their patients.However, some clinical studies have usedmild hypothermia in awake, nonventilatedpatients with ischemic stroke (29–30) oracute myocardial infarction (31), and re-ported cooling awake patients appeared tobe feasible and safe. However, even inthese studies large doses of buspironeand meperidine were used to control shiv-ering, and to allow patients to tolerate thetreatment. Thus, overall, the available evi-dence suggests that proper sedation andanalgesia are important for successful useof induced hypothermia.

Conversely, paralyzing agents have theadvantage that they are highly effective and(in contrast to most sedatives and analge-sics) do not cause hypotension, which canbe an important advantage in some (hemo-dynamically unstable) patients. This is par-ticularly important in the ambulance andemergency room setting where fewer op-tions are available to maintain hemody-namic stability. In this setting, short-termparalysis should be considered more as afirst-line option to control shivering. Thus,the advantages and disadvantages of thedifferent shivering control methods shouldbe carefully weighed in each individual pa-tient. In our view, routine paralysis is usu-ally unnecessary; it seems reasonable to useparalysis only when appropriate sedation/analgesia (and probably magnesium, seelater) have failed to control shivering. Evenin this situation, paralysis is rarely requiredin the maintenance phase, as the shiveringresponse is markedly diminished and oftenceases completely at temperatures below33.5°C (39–41). The sedation strategyshould entail administration of high bolusdoses in the induction phase, and keepingmaintenance doses given via continuousinfusion pumps relatively low. The reasonfor this is that drug clearance changes, andin most cases is markedly reduced, duringmild hypothermia (12, 43–58). Therefore,significant accumulation of drugs includ-ing opiates and sedatives (and muscle par-alyzers, if these are used) is likely to occurwhen high continuous doses are given forprolonged periods during the maintenancephase of moderate hypothermia treatment.

Physiology of Cooling

Heat loss occurs via four basic mech-anisms: convection, conduction, radia-tion, and evaporation (production ofsweat). Under normal circumstances con-

duction (direct transfer of heat from onesurface to an adjacent surface) is negligi-ble, while convective heat loss (transfer ofheat from a surface to the surroundingair) accounts for 20% to 30% of heat lossat room temperature in the absence ofwind. Attempts to induce hypothermia allaim at increasing convective or conduc-tive heat loss. A list of cooling methodsand devices is shown in Table 2. The rateof heat loss is determined by the temper-ature gradient, body composition, andthe conductive properties of the environ-ment. For example, water is a much bet-ter conductor of heat than air, and thuswet skin will transfer heat much moreeasily than dry skin. The rate of heat lossis higher with alcohol-based solutions be-cause the rate of evaporation is higher(Table 2).

The effectiveness of the mechanismscontrolling body temperature decreasewith age; this is due to a decrease insensitivity to small temperature changes(leading to a slower counter-regulatoryresponse), a lower rate of metabolism, aless effective vascular response (i.e., lessvasoconstriction), and frequently a lowerlean body mass index (18, 59). Thus, ingeneral, induction of hypothermia is eas-ier in older patients than in youngerones. Doses of opiates and sedatives re-quired to effectively suppress the body’swarming mechanisms are usually muchhigher in younger patients. Similarly,achieving hypothermia in obese pa-tients can take more time, because ofthe larger mass that needs to be cooledand due to the insulating properties offat tissue.

When initiating hypothermia treat-ment, the treatment period can be di-vided into three phases. The first is theinduction phase, where the aim is to getthe temperature below 34°C and down tothe target temperature as quickly as pos-sible. In this phase a small overshoot(�1°C) should be regarded as acceptableprovided temperature remains �30°C. Inthe maintenance phase the aim should beto tightly control core temperature, withminor or no fluctuations (maximum0.2°C– 0.5°C). Finally, the rewarmingphase should be slow and controlled(warming rate 0.2°C–0.5°C/hr) (74).

Each of the three phases of hypother-mia has specific management problems.In general, the risk of short-term sideeffects, such as hypovolemia, electrolytedisorders, and hyperglycemia, is greatestin the induction phase (1, 12, 61). This isthe phase with the greatest patient “in-


Table 2. Currently available methods and devices for inducing and maintaining hypothermia

Method ManufacturerEfficacy of Induction, General andSpecific Device-Related Advantages

General and Specific Device-RelatedDisadvantages

Surface cooling: airExposure of skin N/A Easy, inexpensive, no procedural risk.

Speed of induction low (�0.5°C/hr)Relatively ineffective. Cannot be used

for maintenance and rewarmingphase

Skin exposure combinedwith water or alcoholsprays or sponge baths

— Easy, inexpensive, relatively effective(alcohol sprays are more effective thanwater sprays). Speed of induction lowto intermediate (�1.0°C/hr)

Labor intensive for nursing staff;patient remains wet for prolongedtimes. Cooling rate cannot be usedfor maintenance and rewarmingphase

Fans Various Easily accomplished, inexpensive Speed ofinduction low to intermediate(�1.0°C/hr)

Additional risks of infection? Cannotbe used for maintenance andrewarming phase

Air-circulating coolingblankets

Polar Air and BairHugger R, ArizantHealthcare, EdenPrairie, MNc,d

Relatively inexpensive; often alreadyavailable due to use in ICU or operatingroom to warm patients. However, speedof induction is very low (�0.5°C/hr)

No more effective than skin exposure(�0.5°C/hr) in the intensive careunit

Specially designed beds Deltatherm R, KCI, SanAntonio, TXd

Cooling rates �1°C/hr. May provide extraprotection from bedsores. Targetedcooling of neck for quicker braincooling. Highly reliable maintenanceand controlled rewarming

Relatively large device, high noiselevels during induction phase. Notmarketed in the United States,device may be withdrawn fromEuropean market also, currentstatus unclear

Surface cooling: fluidsIce packs N/A Easy, inexpensive cooling method. Speed

of induction intermediate (�1°C/hr)Risk of skin lesions and burns. Not

reliable in maintenance andrewarming phase, labor-intensiveif used for maintenance

Complete immersion incold water

N/A Speed of induction excellent (�8°C–10°C/hr);inexpensive

Unpractical, especially in the ICUsetting and for maintaining atarget temperature

Circulating cold waterdirectly against skin ofpatient

LRS ThermoSuit system,Life Recovery Systems,Waldwick, NJc,d

Only tested in animals, clinical trialongoing. Very rapid induction rates(�10°C/hr) in animal study and intheory

Nonreusable; cannot be used formaintenance and rewarming phase

Prerefrigerated cooling pads Laerdal Medi�Cool,Laerdal Medical AS,Stavanger, Norwayd

Polyester and marino wool mesh filledwith polymer crystals. Pads areimmersed in cooled water, dried, andplaced in a freezer for �2 hrs beforeusage. Cooling activity �2 hrs.Preliminary clinical trial ongoing.Low-tech approach, suitable for coolingin the field

No feedback system for temperaturecontrol. Theoretical risk of skininjury. Few data currentlyavailable. Cannot be used forrewarming, difficult to use in themaintenance phase

Prerefrigerated surface pads Emcools AG, Vienna,Austria

Surface cooling elements, tested inanimals, no clinical data yet available.Designed for use in prehospital setting

Cannot be used for maintenance andrewarming phase. Theoretical riskof skin injury. Few data currentlyavailable

Water-circulating coolingblankets

Blanketroll II hyper-hypothermia,Cincinnati Sub-ZeroCompany, Cincinnati,OHc,d

Reusable, significantly lower costscompared to most other devices.Cooling rate �1.0°C–1.5°C/hr with twoblankets

Labor intensive for nursing staffespecially in induction phase. Twoblankets required for quickinduction if used as sole means ofcooling

Blanketroll III hyper-Hypothermia,Cincinnati Sub-ZeroCompany, Cincinnati,OHc,d

Some parts reusable, others disposable.Less labor-intensive than Blanketroll II.Targeted neck cooling possible. Coolingrate �1.5°C/hr. Inexpensive comparedto other disposable devices

Anecdotal evidence suggests somerisk of overshoot in inductionphase

Water-circulating coolingpads

CoolBlue Surface PadSystem, InnercoolTherapies, San Diego,CAc

Disposable vest and thigh wraps blankets.Less labor-intensive than Blanketroll II.Inexpensive compared with otherdisposable devices. Preliminary dataappear promising

Few clinical data available so far;nonreusable materials


stability,” with numerous short-termchanges required in ventilator settings,dosage of vasoactive drugs, insulinpumps, etc. The risks can be minimized

by cooling patients as quickly as possible,i.e., by minimizing the duration of theinduction phase and reaching the morestable maintenance phase as quickly as

possible. This can be accomplished byusing combinations of different coolingmethods, for example, a combination oflarge-volume infusion of cold fluids and

Table 2. —Continued

Method ManufacturerEfficacy of Induction, General andSpecific Device-Related Advantages

General and Specific Device-RelatedDisadvantages

Hydrogel-coatedwater-circulating pads

Arctic Sun TemperatureManagement System,Medivance Inc.,Louisville, COc,d

User friendly, less labor intensive thanwater-circulating blankets; relativelyhigh cooling rates (�1.5°C–2.0°C/hr),with lower percentage of patients’ bodythat needs to be covered to achievecooling. Reliable maintenance andrewarming phase

Slight risk of skin lesions (rednessand mottling) if used at maximumsetting for prolonged time (e.g.,for prolonged fever control)a

Water-circulatingwrapping garments

CritiCool and CureWrapsystems, MTRE,Or-Akiva, Israel; Medi-therm II and IIIsystems, Gaymar,NYc,d

Wraps around the patient; skin contactbetter than with rubber blankets,material is nonadhesive. Cooling rates�1.5°C/hr

Few clinical data available so far;nonreusable materials; causesreversible pressure tracks on theskin

Core coolingIntravascular catheters CoolLine, Coolgard2

and Fortius, AlsiusCorporation, Irvine,CAc,d

Celsius Control System,Innercool therapies,San Diego, CAc,d

SetPoint R and ReprieveRadiant Medical,Redwood City, CAb

Uses intravascular balloons filled withcold saline. Relatively quick inductionrates (�1.5°C–2.0°C/hr), highly reliablemaintenance and rewarming rates.Provides venous access (two sideportsapart from cooling sideports)

Uses metal catheter (10.7F or 14F) toaccelerate heat loss. Quick inductionrates (�2.0°C–4.5°C/hr depending onsize and setting), highly reliablemaintenance and rewarming rates.Venous access via single sideport oninsertion sheath. Temperature sensor incatheter tip for blood temperaturemonitoring

Saline-filled balloons with helix system toimprove heat extraction. Provides rapidcooling rates (�2.0°C–4.5°C/hr),reliable maintenance, and controlledrewarming

Requires invasive procedure, withsome time loss and associatedprocedural risk; single use. Fewdata regarding prolonged usage,preliminary data suggest use up to1 wk is safe. Some risk ofcatheter-related thrombosis

Requires invasive procedure withsome time loss and proceduralrisk. Only one sideport. Few dataregarding prolonged usage. Somerisk of catheter-related thrombosis

Requires invasive procedure withsome time loss and proceduralrisk. No venous access provided bydevice (cooling system only).Some risk of catheter-relatedthrombosis

Infusion of ice-cold (4°C)fluids

N/A Very rapid method to induce hypothermia(�2.5°C–3.5°C/hr); can easily be usedin combination with other methods

Cannot be used to maintaintemperature within narrow range;requires infusion of large volumesof fluid

Peritoneal lavage device Velomedix, SanFrancisco, CA

Automates a cold peritoneal lavage Tested in animals, preliminaryclinical assessments ongoing, notyet commercially available.

Extracorporealcirculation

Various devices Very quick and consistent cooling rates(�4.0°C–6.0°C/hr), reliable

Highly invasive, impractical in theICU setting

Antipyretic agents Various drugs(acetaminophen,aspirin, NSAIDs,others)

Low costs, little additional workload.Cooling rates average �0.1°C–0.6°C

Efficacy is relatively low(temperature decrease�0.1°C–0.6°C) especially inpatients with “central” fever

ICU, intensive care unit; NSAIDS, nonsteroidal anti-inflammatory drugs.aA (low) risk of skin lesions (usually redness and motteling) is probably inherent to all surface cooling devices, but has so far only been reported in one study (71);

bmanufacturer recently went bankrupt, future status of product currently unclear; capproved for clinical use in the United States; dapproved for clinical use in Europe.†The magnitude of this risk is still unclear. This may be similar to “regular” central venous lines, or lower (due to local effects of hypothermia on platelet

function and coagulation), or higher (due to the greater size of these catheters compared to regular central venous catheter).No randomized controlled trial assessing and comparing these devices and methods have been performed. Comments in the table are based on

observational studies, method sections of papers reporting the results of hypothermia trials for various indications, data provided by the manufacturers,data obtained from the Food and Drug Administration website (www.fda.gov), and the experience of the authors.


surface cooling (62). In addition, prelim-inary results suggest that some of thenew intravascular devices may have fastercooling rates than previously used meth-ods. This also applies to rapid infusion ofrefrigerated fluids.

Once the core temperature decreasesbelow �33.5°C the patient tends to be-come more stable, with less risk for fluidloss or intracellular shifts, a cessation orsignificant diminishment of shivering,and a cessation of hypothermia-inducedmajor changes in hemodynamic parame-ters. In this phase, attention should shifttoward the prevention of longer-termside effects such as pneumonia, woundinfections, and bedsores (see later).

Finally, the rewarming phase is asso-ciated with problems of its own, such as(again) electrolyte disorders caused byshifts from the intra-cellular to the extra-cellular compartment. This can be largelyprevented by slow and controlled re-warming. In addition, there are numer-ous animal experiments showing thatrapid rewarming after hypothermia treat-ment can adversely affect outcome (63–65). This is supported by some clinicalobservations. For example, Kawahara etal (66) reported that rapid rewarming fol-lowing cardiac surgery leads to a decreasein jugular venous oxygen saturation, in-dicating hypoxia of the brain; signifi-cantly less jugular desaturations were ob-served when slower rates of rewarmingwere used. Lavinio et al (67) found thatcerebrovascular reactivity was impairedduring rapid rewarming following hypo-thermia treatment. In addition, patientswho developed even mild hyperthermiahad significantly more severe derange-ments of cerebrovascular reactivity, indi-cating the importance of maintainingcontrolled normothermia after hypother-mia treatment (67). Bissonnette et al (68)observed that patients who were rapidlyrewarmed often developed severe brainhyperthermia, even when core tempera-ture measured at other sites remainednormal. Thus, although no studies havebeen performed to determine the opti-mum rewarming rate in cardiac arrestpatients following hypothermia treat-ment, based on the above it is highlyplausible that slow rewarming will betterpreserve hypothermia’s neuroprotectiveeffects.

Cooling methods and devices arelisted and described in Table 2. These canbe broadly divided into (noninvasive) sur-face cooling devices, using adhesive pads,wrapping garments or rubber blankets,

and (invasive) core cooling devices, usingintravascular catheters (made of metal orwith saline-filled balloons).

Most studies using intravascular de-vices have reported highly reliable main-tenance of core temperature, and rela-tively rapid cooling rates once thecatheter is in place (details in Table 2). Adisadvantage is that an insertion proce-dure is required before cooling can beinitiated, and this should be taken intoaccount when calculating the “event-to-target-temperature” time, which is amore clinically relevant end point thanthe cooling rate per se. The time requiredfor insertion strongly depends on theclinical setting (specifically, the rapidavailability at all times of physicians ca-pable of performing the insertion proce-dure) and other logistic factors.

A potential problem is the risk of cath-eter-related thrombosis. Some risk forthrombus formation is inherent to anyindwelling central line (69); ultrasoundstudies have reported central venouscatheter-related thrombus formation in33% to 67% of patients when centralvenous catheter indwelling time was �1week (69). Most of these thrombi remainasymptomatic, and resolve spontaneouslywhen the central line is removed; themain problem is the associated risk ofdeveloping catheter-related infections(69). These issues have not been wellstudied for intravascular cooling devices,so the magnitude of this risk is unknown.Anecdotal evidence from centers usingintravascular cooling devices on a regularbasis suggests that the rate of symptom-atic thrombosis is very low; this may in-crease when a device is left in place forprolonged periods of time. Only onesmall study has used ultrasound to detectthrombus formation associated with anintravascular cooling device, used to in-duce controlled normothermia in TBI pa-tients with an average indwelling time of5 days (70). Although none of these pa-tients had a symptomatic thrombosis, therate of asymptomatic thrombus forma-tion was 50%, with a range of 33% to75% depending on the indwelling time ofthe device. However, the value of thisstudy is limited by its retrospective de-sign and small number of patients (n �11, of which one had to be excluded fromanalysis). Furthermore, the reported rateof asymptomatic catheter-related throm-bosis is similar to what has been observedin studies with “regular” intravasculardevices (69), although it should bepointed out that the catheter indwelling

times were longer in these studies. Thus,it remains unclear whether the risk forcatheter-related thrombosis differs be-tween cooling catheters and “regular”catheters. Clearly, larger and prospectivestudies are needed to properly assess therisks (and efficacy) of these and othercooling devices.

A problem with surface cooling is thatmuch of the patient’s surface area needsto be covered, ranging from 40% to 90%in the induction phase depending on theefficacy of the cooling device and of thecooling pads/blanket. Prolonged and in-tense surface cooling carries a risk of skinlesions. This risk appears to be low, and isrelated to the temperature of the pads/blankets, duration of intense cooling, andthe type of material (higher with rubber,lower with the newer materials). A majoradvantage of surface cooling is that it canbe started immediately, in nurse-drivenprotocols without direct physician inter-vention.

The cooling rates reported in the lit-erature for the older surface cooling tech-nologies are significantly lower than forintravascular catheters and for the newersurface cooling devices (Table 2). Thereare no prospective studies that have com-pared currently available (surface and in-travascular) cooling devices in a stan-dardized way regarding efficacy ofcooling, reliability of maintenance, andrewarming and side effects. Some studieshave attempted to address this issue, butthe results are difficult to interpret be-cause of various methodologic issues.These include enrollment of small num-bers and very different categories of pa-tients; some studies were performed inthe perioperative setting, making the re-sults difficult to compare with studiesenrolling intensive care unit patients. Inthe best comparative study published sofar, Mayer et al (71) assessed the efficacyof two-surface cooling devices for fevermanagement; however, there were prob-lems in the way in which both methodswere applied, in the sense that neitherwas used to maximum efficacy (72). Hoe-demaekers et al (73) compared coolingrates for five invasive and noninvasivedevices in small (n � 10) groups of pa-tients. However, this study enrolled bothpatients requiring controlled hypother-mia and controlled normothermia, mak-ing its results difficult to interpret; asexplained above, achieving controllednormothermia is more difficult, and re-quires a different approach, than con-trolled hypothermia. Furthermore, some


aspects of the methods employed to cal-culate cooling rates were unclear in thisstudy.

Significant increases in cooling ratescan be achieved by using cold fluid infu-sion as an accessory cooling method (62).This implies that cooling devices shouldnot only be judged solely or mainly onthe basis of their cooling speed, but also(and perhaps especially) on their reliabil-ity to maintain target temperature withina narrow range and to slowly and safelyrewarm the patient, as well the device-associated side effects and device-relatedphysician and nursing workload (74).

Side Effects ofInduced Hypothermia

Hypothermia induces physiologicchanges in virtually every organ of thebody. The kinetic properties of most en-zyme systems are temperature-depen-dent; thus, the speed of various enzyme-mediated reactions is significantlyinfluenced by hypothermia. This means,for example, that drug metabolism is sig-nificantly affected by induction of hypo-thermia (43). It should be realized thatthe distinction between physiologic(“normal”) consequences and genuineside effects of hypothermia is to someextent artificial; some changes are phys-iologic, but are nevertheless undesirablein critically ill patients and therefore re-quire preventive measures and/or proac-tive treatment. In contrast, other conse-quences of hypothermia can be regardedas side effects, but pose no great risk tothe patient and thus usually do not re-quire active treatment.

The most important side effects andchanges are listed in Table 3. Details arediscussed below.

Arrhythmias, Hemodynamic Changes,and Cardiovascular Effects. During mild-to-moderate hypothermia (32°C–34°C),cardiac output decreases by 25% to 40%mainly due to a decrease in heart rate(see later). In general, the decrease inmetabolic rate is equal to or greater thanthe decrease in cardiac output. Tempera-ture-corrected (mixed) venous saturationincreases slightly or remains unchanged,reflecting an unchanged or improved cir-culatory state. Central venous pressureusually rises, and there is also an increasein arterial resistance (systemic vascularresistance) and a slight rise in blood pres-sure (by �10 mm Hg) due to hypother-mia-induced vasoconstriction of periph-eral arteries and arterioles (75, 76). This

vasoconstrictive effect is absent or lesspronounced in cerebral arteries, wherethe balance between cerebral blood flowand cerebral metabolism is maintained orimproved (77–81). In theory, an increasein SVR could increase the workload of theinjured heart by increasing afterload.However, most patients who have had acirculatory collapse followed by reperfu-sion develop a systemic inflammatory re-sponse syndrome-like response with apathologic decrease in vascular tone. Inthis situation an increase in systemic vas-cular resistance and vascular tone will bebeneficial, and will also increase coronaryperfusion. Furthermore, the hypother-mia-induced decrease in heart rate (seelater) will decrease myocardial oxygen de-mand.

The effect of hypothermia on myocar-dial contractility is strongly dependent onheart rate. If the heart rate is allowed todecrease along with the temperature,myocardial contractility as measured bysystolic function usually increases, al-though there may be a mild degree ofdiastolic dysfunction (82–85). However,if the heart rate is artificially increasedthrough administration of chronotropicdrugs or a pacing wire, myocardial con-tractility decreases significantly. Thisphenomenon has been demonstrated inanimal studies and also in patients un-dergoing cardiothoracic surgery (86).Thus, the effect of hypothermia on myo-cardial function strongly depends onwhether the heart rate is allowed to de-crease.

The fact that hypothermia can indeedimprove circulatory parameters is shownby its successful usage in five small stud-ies (three in pediatric patients and two inadults) to improve circulation and re-verse refractory cardiac shock followingcardiac surgery (87–91). Two small stud-ies have reported that hypothermia canbe used safely and effectively in hemody-namically unstable patients who remaincomatose following cardiac arrest (166,167). Whether hypothermia improves he-modynamic stability also depends on pre-vention of hypovolemia, which can de-velop because of hypothermia-induced“cold diuresis.” If hemodynamic instabil-ity develops in the induction phase ofcooling, hypovolemia is the most likelycause and a fluid challenge test is war-ranted.

Hypothermia also induces electrocar-diographic changes and alterations in theheart rhythm. When hypothermic treat-

ment is initiated and body temperaturebegins to drop, mild sinus tachycardiawill initially develop. This is partly due toan increase in the venous return to theheart caused by a shift in circulatory vol-ume from the peripheral compartment(especially the skin) to the core compart-ment, leading to a reflex increase in heartrate. Sinus bradycardia ensues as temper-ature drops below 35.5°C, with a progres-sive decrease in heart rate as temperaturedecreases further. At core temperaturesof �32°C the heart rate typically de-creases to around 40–45 beats/min oreven lower, although there is wide inter-and intraindividual variability and heartrate may remain at �60 or even higher(1). This phenomenon is caused by a de-crease in the rate of diastolic repolarizationin the cells of the sinus node. Electrocar-diographic changes include prolonged PRinterval, widening of the QRS complex, in-creased QT interval, and sometimes so-called Osborne waves (Fig. 1). Thesechanges usually do not require treatment;as explained above, at 32°C a heart rate of40 is perfectly normal. If a stimulation ofheart rate is deemed necessary, this can beaccomplished by administering isoprenalinor dopamine, by rewarming the patient toslightly higher temperatures or (in extremecases) by inserting a pacing wire. Atropineis ineffective in this situation. It should bekept in mind that excessive stimulation ofheart rate during hypothermia is likely todecrease myocardial contractility (86), andthat artificially increasing the heart rate byadministration of chronotropic medication israrely necessary. Conversely, if the heart ratedoes not decrease during hypothermia, careshould be taken to rule out insufficient seda-tion as a cause of the (relative) tachycardia.

Whether or not the heart rate (or,more correctly, cardiac output) is suffi-cient in an individual patient can be de-termined by, for example, measuring thetemperature-corrected (mixed) venoussaturation and/or changes in metabolicparameters such as lactate levels. Regard-ing the latter parameter, it should berealized that lactate levels frequently in-crease during hypothermia (usually toaround mmol/L with maximum of 5–6mmol/L); however, once the target tem-perature has been reached these levelsshould remain stable. If lactate and met-abolic acidosis increase further this mayindicate insufficient circulation, requir-ing further diagnostic evaluation and per-haps therapeutic interventions such asfluid administration and/or administra-tion of inotropic drugs.


Table 3. The most important physiological changes and potential side effects of hypothermia

Effect Frequency Treatment Required Comment

Hypovolemia (due to cold diuresis);hypotension due to hypovolemia

Intermediate Yes More frequent in TBI than cardiac arrest, probablydue to concomitant treatments such asmannitol administration

Cardiovascular changes: 1bloodpressure, CVP, and mixed venoussaturation; 2heart rate, 2Cardiacoutput

Frequent Usually no Mean arterial pressure increases slightly (�10 mmHg) during mild hypothermia. Cardiac outputdecreases but at a rate equal to or less thandecrease in metabolic rate. The net result isunchanged or improved balance between supplyand demand

EKG changes: bradycardia, 1PR intervaland QT interval, �QRS complex.Arrhythmias can develop at temp�28°C–30°C

Frequent Usually no There is no hypothermia-induced risk for severearrhythmias unless core temperature decreasesto �30°C. The risk increases significantly attemperature �28°C. Initial type of arrhythmiais usually artrial fibrillation; other arrhythmiasincluding ventricular fibrillation may follow

Electrolyte disordersa (loss of K, Mg, P,Ca); in rewarming phase risk ofhyperkalemia due to extracellular shift

Frequent Yes Far more frequent in TBI due to concomitanttreatments such as mannitol administration.Maintain electrolyte levels in the high normalrange (Mg �1.0 mmol/L, K �4.0, and PO �1.0mmol/L, respectively) in all patients duringhypothermia treatment. Rewarm slowly to avoidhyperkalemia in rewarming phase

Impaired coagulation cascade, risk ofbleeding

Mild impairment ofcoagulation:frequent.Bleeding: Rare

Usually no Platelet count and function and coagulation areimpaired during hypothermia but bleedingproblems are rare. No hypothermia study hasreported significant problems with bleeding.Consider platelet administration before surgery/invasive procedures. Administration ofDesmopressin may improve platelet functionduring cooling (162)

Shivering Frequent Yes Shivering leads to rewarming and should becontrolled using intravenous magnesium,analgesia (bolus dose of meperidine or quick-acting opiates), sedation (propofol,benzodiazepines), and if necessary brief-actingparalytic drugs. Other drugs with antishiveringeffects: clonidine, ketanserin, tramadol, urapidil,and doxapram. The cardiovascular effects duringshivering are different than in the post-operative setting. Shivering response issignificantly blunted when temperaturedecreases below �33.5°C

Increased infection risk, particularlyairway and wound infections

Intermediate Yes Inflammatory response is suppressed by cooling.Prolonged (�24 hrs) cooling is associated withhigher infection risk. Consider antibioticprophylaxis

Insulin resistance, hyperglycemia Frequent Yes Insulin doses required to maintain normoglycemiamay increase during induction, and decreaseduring rewarming

Bedsores (due to vasoconstriction in theskin, immobilization, and immunesuppression)

Intermediate Yes Extra attention required for bedsore preventiondue to converging of risk factors

Lab changes: 1amylase, liver enzymes,lactate, ketonic acids, and glycerol;2white blood cells and platelets; mild1hematocrit; mild acidosis

Frequent No Usually no interventions required. Some lab values(coagulation parameters, blood gases, pH value)are influenced by temperature and should betemperature-corrected

Changes (usually decrease) in drugclearance (due to slowing ofnumerous liver enzymes)

Frequent Yes Adjust infusion rates; use bolus doses rather thanincreasing continuous infusion

CVP, central venous pressure; EKG, electrocardiogram; TBI, trumatic brain injury.aThe magnitude of this risk is still unclear. This may be similar to “regular” central venous lines, or lower (due to local effects of hypothermia on platelet

function and coagulation), or higher (due to the greater size of these catheters compared to regular central venous catheters).


The risk of developing clinically sig-nificant arrhythmias due to hypothermiais extremely low as long as the patients’core temperature remains higher than30°C. In fact, experimental evidence sug-gests that mild hypothermia may in-crease membrane stability, thereby de-creasing the risk of arrhythmias (175).Animal studies have shown that mild tomoderate hypothermia significantly im-proves the likelihood of successful defi-brillation and good resuscition outcomes(175–177). Various case reports describesuccessful use of hypothermia to treatarrhythmias (junctional ectopic tachycar-dia) in young infants (178–181), indicat-ing that mild hypothermia indeed canimprove membrane stability and decreasethe likelihood of arrhythymias.

In contrast, more profound hypother-mia (�28°C, in rare cases 30°C if electro-lyte disorders and/or ischemia arepresent) can increase the risk of arrhyth-mias. This usually begins with atrial fi-brillation, which can be followed by moresevere arrhythmias including ventricularfibrillation (VF) and ventricular tachycar-dia, if temperature decreases further. It isimportant to realize that mechanical ma-nipulations can trigger this transitionfrom atrial fibrillation to VF, as the myo-cardium becomes highly sensitive to suchmanipulations during hypothermia (92).Thus, if a physician decides to performchest compressions because of extreme

bradycardia this can easily lead to a con-version of sinus rhythm to VF, or atrialfibrillation to VF. A major problem is thatonce arrhythmias do develop in deeplyhypothermic patients these are far moredifficult to treat than at mild hypother-mia or normothermia. The reason forthis is that the myocardium at deep hy-pothermia becomes less responsive tomany anti-arrhythmic drugs (93–96). Asthese problems are rarely encountered attemperatures above 28°C–30°C, greatcare should be taken to keep core tem-peratures above this limit.

The increase in venous return inducedby hypothermia can lead to activation ofatrial natriuretic peptide and a decreasein the levels of anti-diuretic hormone.This (in combination with other mecha-nisms such as tubular dysfunction, seelater) can lead to a marked increase indiuresis (“cold diuresis”), which if uncor-rected can lead to hypovolemia, renalelectrolyte loss, and hemoconcentrationwith increased blood viscosity (61, 97–98). The risk of hypovolemia is greater ifthe patient is simultaneously treated withagents that can increase diuresis, such asmannitol in TBI patients. The increase inblood viscosity (�2% per °C decrease incore temperature) can lead to problemsin the microcirculation; the above-mentioned mechanisms combined withtubular dysfunction can lead to severeelectrolyte disorders (see later), including

a rise in serum sodium levels and osmo-larity. Thus, careful attention should bepaid to intravascular volume and fluidbalance in patients treated with hypo-thermia, and hypovolemia should beavoided or promptly corrected.

Several studies have shown that mildhypothermia in awake patients leads toan increase in plasma levels of cat-echolamines (169, 170). However, studiesin hypothermic patients under generalanesthesia found no such difference incatecholamine levels. Indeed one studyreported lower catecholamine levels inmildy hypothermic patients comparedwith controls (171). Conflicting observa-tions have been reported from studies inchildren undergoing cardiac surgical pro-cedures; one study reported increasedcatecholamine levels when patients werecooled to 26°C (172); however, anotherstudy reported no significant changes incatecolamine levels during surface cool-ing to 28°C, nor during circulatory arrestand cooling until 18°C (173). However,catecholamine levels did increase signifi-cantly during rewarming (172). Thus,catecholamine levels may increase duringmild-to-moderate hypothermia, but thisphenomenon may be linked to a shiver-ing/stress response and to rewarmingrather than to hypothermia per se. Inaddition, some animal studies suggestthat hypothermia reduces ischemia-induced catecholamine release from hy-poperfused tissues, while catecholaminesecretion from noninjured tissues mayincrease (174). Thus, the relationship be-tween hypothermia and catecholaminelevels is a complicated one, which may beinfluenced by concomitant drug admin-istration (especially sedatives) and otherfactors.

Coronary Perfusion and Ischemia. Asexplained above, perioperative hypother-mia is associated with an increased risk ofmorbid cardiac events (15–17). Partly dueto these observations, it is widely believedthat hypothermia can cause coronary va-soconstriction and myocardial ischemia.However, the real situation is more com-plex. In healthy subjects, mild hypother-mia (�35°C) actually increases coronaryperfusion (100, 169). However, this is lessclear in patients with coronary artery dis-ease; hypothermia can induce vasocon-striction in severely atherosclerotic coro-nary arteries (100). Thus, which effectwill occur may depend on the preexistinghealth of the patients’ coronary arteriesand on local factors.

Figure 1. Electrocardiogram (EKG) changes during hypothermia treatment. In this patient, many ofthe EKG changes associated with hypothermia are apparent: Note the slightly prolonged PR interval,slight widening of the QRS complex and increased QT interval and Osborn wave (arrow). In this patientthe changes developed during the induction phase of cooling persisted throughout the hypothermiatreatment and disappeared at the end of the slow rewarming process. The EKG at a core temperatureof 35.9°C.


On the basis of these observations onemight reasonably expect hypothermia-induced coronary vasoconstriction to oc-cur in patients who have been admittedfollowing a cardiac arrest, based on thepresence of coronary artery disease whichhas caused the cardiac arrest in the firstplace. However, as outlined above, vari-ous animal experiments (101–108) andpreliminary clinical studies (109, 110)have shown that hypothermia may actu-ally decrease myocardial injury followingcardiac arrest, provided it is initiatedearly enough. Studies in patients under-going cardiac surgical procedures havereported that under hypothermic condi-tions the reduction in cardiac work wasgreater than the reduction in coronaryblood flow (111, 112). Thus, although thequestion whether hypothermia mitigatesmyocardial injury needs to be addressedin further (larger) studies, the availableevidence certainly does not suggest thathypothermia increases such injury throughhypothermia-induced vasoconstriction.

Drug Clearance. As outlined above,not only the serum levels and drug clear-ance but also the effects of various drugsmay change (43–58). An example of thelatter is the response to vasoactive drugssuch as adrenalin and noradrenalin,which may be slightly blunted by hypo-thermia (52). Conversely, the half-life ofvasopressors is increased, so higher con-centrations will be achieved with thesame dose. Apart from vasoactive pressordrugs, effects of hypothermia on druglevels and/or drug actions have beendemonstrated for the following drugs:fentanyl, remifentanyl, and morphine;propofol, barbiturates, and midazolam;neuromuscular blocking agents such asvecuronium, rocuronium, and atra-curium; phenytoin; nitrates; propanolol;and tissue/gas partition coefficients ofvolatile anesthetics (43–58). In mostcases the effect of hypothermia is to in-crease drug levels and/or enhance the ef-fect of the drug. The underlying mecha-nism is a reduction in the activity ofmany liver enzymes during hypothermia,combined with reduced perfusion of theliver and reduced production of bile lead-ing to decreased excretion of some drugs.Changes in distribution volume and hy-pothermia-induced tubular dysfunctionmay also play a role. It seems likely thatthe metabolism of other drugs will beaffected by temperature in a similar way,based on their excretion mechanism.These mechanisms should be taken intoaccount when treating patients under hy-

pothermic conditions. Sedation and anal-gesia should be a specific focus of atten-tion; benzodiazepines and opiates,particularly morphine (58), can accumu-late during hypothermia, complicatingneurologic assessment after treatment. Asexplained above, adequate sedation is of keyimportance for effective cooling, but cor-rect dosing can be difficult under hypother-mic conditions.

Electrolyte Disorders. Electrolyte dis-orders may develop especially in the in-duction phase of cooling. Patients fre-quently develop low electrolyte levelsduring cooling; the reason is a combina-tion of increased renal excretion (causedby a combination of cold diuresis andtubular dysfunction) and intracellularshift (61). Such electrolyte disorders canincrease the risk of arrhythmias and haveother adverse effects on outcome. Magne-sium (Mg), in particular, may play animportant role in mitigating brain inju-ries, myocardial injury, and arrhythmias(113–121). Mg depletion significantly in-creases brain injury in various animalmodels for TBI and stroke, and clinicalstudies have shown improved neurologicoutcome with Mg supplementation in se-vere eclampsia and subarachnoid hemor-rhage (114–116). Hypomagnesaemia isalso associated with adverse outcome inpatients with unstable angina or myocar-dial infarction (117, 118). The latter issueis particularly relevant if patients aretreated with induced hypothermia follow-ing cardiac arrest. Several studies havelinked hypomagnesemia to increasedmortality in the intensive care unit (119,120) and in the general ward (121).

All this implies that levels of Mgshould be maintained in the high-normal�but perhaps not supranormal (122)�range in patients with neurologic injuriesin general, and those treated with hypo-thermia, in particular. This also applies toother electrolytes such as potassium andphosphate. Potassium levels may riseduring the rewarming phase, as potas-sium that was secreted into the cell in theinduction phase is released. This is one ofthe reasons why rewarming should bedone very slowly, giving the kidneys timeto excrete the excess potassium. Hyperka-lemia will not develop if rewarming is slowand if renal function is not grossly impaired(61).

Hyperglycemia. As explained abovehypothermia can simultaneously de-crease insulin sensitivity and reduce in-sulin secretion by pancreatic islet cells.Thus, patients treated with induced hy-

pothermia will be at higher-than-averagerisk for developing hyperglycemia, andhyperglycemia may become more severe(or insulin requirements may increase)during cooling (1). This requires activemanagement because hyperglycemia mayadversely affect outcome in critically illpatients (123, 124). In addition, hypergly-cemia may have specific negative effectson neurologic outcome, for example byincreasing brain injury during episodes ofischemia (125–128). Thus, despite thefact that it is a physiologic consequenceof hypothermia, hyperglycemia should behigh on the list of potentially preventableside effects. What the appropriate targetvalue for blood glucose should be is cur-rently unclear, as there is conflicting ev-idence regarding the risks and benefits ofvery tight glucose control in some cate-gories of patients (123, 124, 129–131); amajor study addressing this issue is cur-rently ongoing (132). However, it seemsprudent to avoid at least severe hypergly-cemia during cooling, with target valuesof 4–8 mmol/L.

Other Metabolic Effects and BloodGas Management. Hypothermia leads toan increase in the synthesis of glycerol,free fatty acids, ketonic acids, and lactate,causing a mild metabolic acidosis, whichin most patients does not require treat-ment. In contrast to the pH levels mea-sured extracellularly, intracellular pHlevels increase slightly during cooling.

The hypothermia-induced decrease inmetabolic rate (�8% per °C drop in coretemperature) also reduces oxygen con-sumption and CO2 production. Ventilatorsettings should be adjusted during induc-tion of hypothermia, and blood gasesshould be monitored frequently espe-cially during the induction phase. Itshould also be realized that blood gasvalues are temperature dependent. Be-cause blood gas analyzers warm bloodsamples to a temperature of 37°C beforeanalysis, when the actual temperaturediffers significantly from 37°C these mea-surements will not be correct; in bloodsamples from hypothermic patients PO2

and PCO2 will be overestimated, and pHunderestimated. For example, in a patientwith a core temperature of 30°C and PCO2

determined to be 40 mm Hg in an uncor-rected measurement, the temperature-corrected PCO2 value would be 29 mm Hg(133). In the same patient with an uncor-rected PO2 of 100 mm Hg, the tempera-ture-corrected value would be 73 mm Hg.By the same token true pH is underesti-mated, with hypothermia leading to a


more alkalotic status compared with nor-mothermia (pH increase � 0.012 pHunits/°C). In the example cited above(measured CO2 40, corrected CO2 29) themeasured pH would be 7.40, whereas thetemperature-corrected pH would be 7.50.

The concept of CO2 management inwhich the PCO2 obtained by measurementat 37°C is kept constant (for example, at40 mm Hg) regardless of the actual bodytemperature is called alpha-stat. If thePCO2 value is corrected for the actualbody temperature, and is held constant atthe same level as during normothermia,this implies that the “true” amount ofCO2 will increase during hypothermia.This concept of CO2 management iscalled pH-stat. In other words, when al-pha-stat CO2 management is applied, pHthat is not corrected for current bodytemperature remains constant, whiletrue pH increases because the actualPaCO2 has decreased. When pH-stat CO2

management is used, true pH remainsconstant, while pH that is not correctedfor temperature decreases. The effects oftemperature on blood gas managementhave been studied most extensively in thecontext of hypothermic coronary andlarge vessel surgery, where the differ-ences can be much more pronounced be-cause the temperatures used are oftenlower than the 32°C–34°C range used inthe intensive care unit (134–136).

The issue of which method is superiorin the management of hypothermic pa-tients has not been settled, and will notbe extensively discussed here. Supportingevidence can be found for both methods(133–138). Supporters of the pH-statmethod argue that application of alpha-stat management leads to hyperventila-tion, hypocapnia, and hypocapnia-in-duced cerebral vasoconstriction, whilepH-stat management induces a degree ofhypercapnia leading to cerebral vasodila-tion (provided cerebral autoregulation isintact); the latter would seem to be amore attractive option. However, hyper-capnia can impair or abolish cerebral au-toregulation, presumably because CO2-induced vasodilatation limits the vessels’capacity to dilate further. This could im-ply that the body’s capacity to divertblood to injured areas would be impaired.In addition, some animal studies suggestthat respiratory alkalosis is physiologi-cally appropriate during hypothermia topreserve normal physiologic conditions(135). In our clinic, we use a modifiedalpha-stat method where gases are tem-perature-corrected for CO2 but slightly

below-normal values (32–34 mm Hg,which is 42–44 mm Hg at 37°C) aremaintained during hypothermia.

Regardless of which method of bloodgas management is chosen, physicians us-ing mild hypothermia in their patientsshould be aware of the effects of tempera-ture on blood gas analysis (as well as onother laboratory measurements such as co-agulation parameters). In addition, regard-less of whether the alpha-stat or pH-statmethod is used to guide pH/CO2 values, theeffect of temperature on PO2 should alwaysbe taken into account, and PaO2 shouldalways be corrected for actual current bodytemperature in hypothermic patients.Thus, it would be a mistake to adapt in-spired oxygen fraction to the apparentlyhigh uncorrected values of PaO2 obtainedduring hypothermia. This also applies tomeasurements of mixed venous or venoussaturation, which will be influenced by coretemperature in the same way. If it is notpossible to obtain blood gas results mea-sured at the patient’s true core tempera-ture, values can be estimated using thefollowing rule of thumb: for PO2, subtract 5mm Hg for every 1°C below 37°C; for PCO2,subtract 2 mm Hg for every 1°C below37°C; for pH, add 0.012 points for every 1°Cbelow 37°C.

Coagulation Parameters. Hypother-mia induces a mild bleeding diathesis,with increased bleeding time due to ef-fects on platelet count, platelet function,the kinetics of clotting enzymes and plas-minogen activator inhibitors, and othersteps in the coagulation cascade (139–145). Standard coagulation tests willshow no abnormalities unless they areperformed at the patient’s actual coretemperature; as is the case with blood gasanalyses usually the samples are warmedto 37°C before the clotting tests are per-formed. Despite the coagulation defectsthat can be caused by hypothermia, therisk of clinically significant bleeding in-duced by hypothermia in patients whoare not already actively bleeding is verylow. None of the large clinical trials inpatients with TBI, subarachnoid hemor-rhage, stroke, or postanoxic coma hasreported significantly increased risks ofbleeding associated with hypothermia.The situation may be different in patientswho are already actively bleeding, for ex-ample in multitraumatized patients. Inthis situation the sites of bleeding shouldbe controlled before hypothermic therapyis initiated. Of note, hypothermia doesnot begin to affect platelet function untiltemperature decreases below 35°C (139–

145); clotting factors are affected onlywhen temperature decreases below 33°C(140–142, 144). Thus, very mild hypo-thermia of 35°C does not affect clottingin any way, and this may be the temper-ature of choice for patients who have anindication for therapeutic hypothermiabut are actively bleeding, or who are atvery high risk for bleeding.

Infections. Hypothermia impairs im-mune functions and inhibits various in-flammatory responses. Indeed, this sideeffect is inherent to the treatment be-cause impairment of harmful inflamma-tory reactions in the brain may be one ofthe mechanisms through which hypo-thermia can exert protective effects (1, 3).Hypothermia inhibits the secretion ofproinflammatory cytokines and suppres-sion leukocyte migration and phagocyto-sis (1, 3, 146). Hypothermia-induced in-sulin resistance and hyperglycemia mayfurther increase infection risks. Some ofthe clinical studies using induced hypo-thermia for various indications have re-ported a slightly, moderately and, in somecases, severely increased incidence of pneu-monia when hypothermia was used for pe-riods longer than 24 hours. However, moststudies using hypothermia for 24 hours orless have reported no or only small in-creases in the infection rates. Antibioticprophylaxis in the form of selective decon-tamination of the digestive tract can reduceGram-negative infection rates and perhapsreduce mortality (147); there is some evi-dence that selective decontamination of thedigestive tract can be used to prevent infec-tions during prolonged use of hypothermia(98, 148), and depending on the setting thiscould be considered for patients undergo-ing hypothermia treatment.

Hypothermia also increases the risk ofwound infections (149). This may be re-lated both to diminished leukocyte func-tion and to hypothermia-induced vaso-constriction in the skin. Thus extra careshould be taken in cooled patients to pre-vent bedsores, which are more likely toshow progression and/or impaired heal-ing. In addition, extra attention should bepaid to catheter insertion sites and to anysurgical wounds which may be present.

Shivering. The problems and meta-bolic consequences of shivering in theinduction phase have been discussedabove. Shivering can be counteracted byadministration of sedatives, anesthetics,opiates, and/or paralyzing drugs (Table4). In most patients, shivering can besignificantly attenuated by relativelysmall doses of opiates. When using para-


Table 4. Drugs that can be used to control shivering

Drug EfficacyHypotensive

EffectSedativeEffecta Additional Comments, Advantages, and Disadvantages

Magnesium (2–3 g)b �� Advantages: some evidence for direct neuroprotectiveeffects of Mg. “Pre-emptive” correction ofhypothermia-induced Mg depletion

Propofol (20–150 mg)b �� Advantages: brief-acting. Anti-seizure effect.Disadvantage: more pronounced hypotension

Benzodiazepines (dosedepending on typeof drug; e.g.midazolam 2.5–10mg)b

�� Advantages: less hypotension. Disadvantages:Complicates neurological evaluation. Reducedmetabolism during cooling can lead to drugaccumulation with persistent sedative effect afterrewarming

Meperidine 10–25 mg �� Advantages: rapid (1–5 mins) effect. Effect lastslonger than with quick-acting opioids. Effect morepronounced than other opioids because of activityat kappa receptors. Relatively mild hypotensiveeffect. Disadvantages: complicates neurologicalevaluation. Slower metabolism during cooling.

Quick-acting opiates:fentanyl 50–100g,b alfentanyl100–250 g

�� Advantages: rapid (1–5 mins) effect. Relatively mildhypotensive effect. Disadvantages: complicatesneurological evaluation. Decreased drugmetabolism during cooling

Morphine 2.5–5 mgb �� Advantage: low costs; additional sedative effect.Disadvantages: delayed (20 mins) effect. Greaterhypotensive effect compared with fentanyl

Dexmedetomidine50–100 gb

�� Advantages: brief-acting; only mild hypotension.Disadvantages: only moderately effective;expensive. Currently not available in Europe

Clonidine 75–200 gb �� Effect in 4–7 mins. Disadvantages: Hypotension,additional bradycardia

Ketanserin 10 mgb �� Effect in 4–7 mins. Advantages: increases coolingrate. Disadvantage: moderate hypotensive effect

Tramadol 50–100 mg �� More rapid effect than morphine (�5 mins).Relatively mild hypotensive effect. Disadvantages:complicates neurological evaluation. Metabolismdecreases during hypothermia. Can cause seizures

Urapidil 10–20 mg ��? �� Conflicting results of studies on efficacy.Disadvantage: pronounced hypotensive effect

Doxapram 100 mg �� Advantages: rapid action (1–5 mins). Can increaseheart rate and blood pressure. Disadvantages: cancause laryngeal spasms

Physostigmine 2 mg �� Can cause additional bradycardia and hypotensionFlumazenil 0.25–0.5

mg�� Few data available. Efficacy may be lower outside the

perioperative settingNefopam 10–20 mg �� Can induce convulsions and anaphylactic reactions.

Currently not available in the United StatesMetamizol � Low efficacyOndansetron � � Low efficacyOther options:

lidocaine,nalbuphine,pentazocine,methylphenidate

/� Questionable or no efficacy

Muscle paralyzers �� Advantage: 100% effective. Disadvantages: does notaffect neurological triggers for shivering; maymask insufficient sedation and/or seizure activity;long-term risks of critical illness neuropathy

For most drugs efficacy increases at higher doses. Efficacy scale: , not effective; �, somewhat effective; ��, moderately effective; ��, effective;��, highly effective; ��, 100% effective. Side effect scale: , no risk; �, mild risk; ��, moderate risk; ��, clear risk; ��, high risk.

aHaving a sedative effect can be both advantageous (suppression of stress response, vasodilation with increased heat loss) and disadvantageous(complication of neurological evaluation, hypotension). Reduced metabolism during cooling can lead to drug accumulation with persistent sedative effectafter rewarming; this applies especially to longer-acting drugs such as benzodiazepines and morphine, where a persistent sedative effect can complicateneurological evaluation; bcan also be given as continuous infusion. General methods: there is some evidence that warm-air skin counterwarming can beused to combat shivering. Drugs such as acetamoinophen (paracetamol), buspirone, and/or nonsteroidal anti-inflammatory drugs can be used to lower theshivering threshold.


lyzing agents and/or when opiates aredeemed undesirable, alternatives to treatshivering include administration ofclonidine, neostigmine, and ketanserin.However, care should be taken to avoid ad-verse effects; for example, clonidine mayworsen hypothermia-induced bradycardia.Warming of the hands, feet, and face (“skincounterwarming”) may also reduce the shiv-ering response (54, 60, 153–155); combina-tion with sedation may be required to achievethis effect (99, 156).

Other Side Effects. Hypothermia is as-sociated with impaired bowel functionand may aggravate gastric emptyingproblems. In addition, a myriad ofchanges in laboratory measurements canoccur; apart from hyperglycemia andelectrolyte disorders, the most frequentlyoccurring changes are a rise in liver en-zymes and serum amylase, mild increasein serum lactate levels (average, 2.5–5mmol/L; but may be higher in somecases) as well as ketonic acids and glyc-erol (leading to a mild metabolic acido-sis), and a decrease in platelet count andsometimes white blood cell count (12).

Monitoring Temperature andGuiding Hypothermia Treatment

When applying induced hypothermia,it is of key importance that core temper-ature be measured accurately, and thatthe site chosen to measure core temper-ature reflect “true” core temperature. Al-though in most patients the goal of treat-ment is to lower brain temperature, sideeffects will be determined mainly by thetemperatures in other organs. The gener-ally accepted gold standard for “true”core temperature is the temperature ofthe blood, measured with a pulmonaryartery catheter (151). All other sites fortemperature measurement should becompared with this gold standard.

Most cooling devices now work with acontrolled feedback system that continu-ously measures a patient’s temperatureand changes the temperature of the cool-ing element (catheters, pads, or blankets)accordingly.

In this regard, it should be realizedthat most of the devices and probes thatare currently used to monitor core tem-perature in critically ill patients were notdesigned to detect rapid changes in tem-perature; rather, they were designed toreflect small temperature changes overprolonged periods of time as accurately aspossible. The probes take some time toequilibrate, and monitoring devices to

which the probes are connected usuallyhave a low sampling frequency of temper-ature readings leading to further delaysin registering temperature changes.

Many of the new cooling devices haverelatively high cooling rates; especiallywhen combined with cold fluid infusion,cooling rates of 4°C/hr or more can beachieved. Such rapid cooling will inevita-bly lead to a time lag between registeredtemperature and measured core temper-ature, unless blood temperature is mea-sured directly. This applies to all of themost commonly used core temperaturemonitoring sites (bladder, rectum, esoph-agus, and tympanic membrane), al-though the lag times differ considerably.In the induction phase this time lag canlead to a significant “overshoot” of coretemperature below the desired target, asthe cooling device continues cooling(based on the measured temperature)while the target temperature has in factalready been reached. The faster the cool-ing rates, the greater will be the time lagbetween registered and actual core tem-perature. This problem can be avoided byusing blood temperature to control cool-ing rates. A pulmonary artery cathetercould be used for this purpose, and somecooling catheters have an inbuilt temper-ature sensor at the catheter tip to guidetherapy.

The equilibration rate between theblood and the organ where temperatureis measured is influenced by a number offactors including the type of organ, organperfusion (which in turn is influenced bysystemic factors such as presence ofshock, hypotension, or hypovolemia—with slower equilibration when perfusionis decreased—and by preexisting diseasesuch as atherosclerosis, again with slowerequilibration if severe atherosclerosis ispresent), and by various local factors. Thecooling method employed (especially sur-face vs. invasive cooling) can also affecttemperature readings; invasive coolingmethods cool the blood directly andtherefore will affect temperature readingsmore rapidly.

Some technical issues may furthercompound this problem. For example,many of the bladder temperature probes(Foley catheters with temperature sen-sors) that are widely used (e.g., Covidien,Mansfield, MA) have a “floating” temper-ature sensor that is not welded to theinterior of the catheter. The sensor cantherefore move within the catheter, and iftraction is applied to the catheter the tipof the temperature probe can move back-

ward inside the saline-filled balloon thatoccludes the bladder. If the balloon hasjust been filled with room-temperaturesaline, this can significantly affect tem-perature readings.

For all these reasons, during the in-duction phase the registered temperaturewill constantly lag behind the actual coretemperature. This means that the coolingdevice may continue to cool the patienteven when target temperature has al-ready been reached, because the temper-ature reading of the probe controlling thecooling device lags behind the actual coretemperature. The device will continuecooling until the temperature input ap-proaches the set target temperature; bythis time the actual core temperature (asmeasured in the blood) may have droppedsignificantly below the target range.

All of the most commonly used tem-perature monitoring sites have specificadvantages and problems. The averagetime lags between various core monitor-ing sites and the blood, as well as specificadvantages and limitations of variouscore temperature monitoring sites, aresummarized in Table 5.

Fever Control

Controlled normothermia representsa closely related but separate area of ther-apeutic temperature management. Inmany aspects inducing and maintainingnormothermia is less problematic thaninducing hypothermia, because the num-ber of potential side effects is muchlower. For example, side effects such assuppression of immune function and co-agulation disorders do not occur duringcontrolled normothermia (although theproinflammatory state associated with afever response can be blunted by preven-tion of fever). In contrast, side effectssuch as shivering may be much morepronounced during controlled normo-thermia than in induced hypothermia.The reason for this is that the body’scounter-regulatory mechanisms decreasesignificantly at lower temperatures, butwork at maximum efficiency in the nor-mal range. In patients with fever, thehypothalamic setpoint that regulates coretemperature is temporarily “reset” to ahigher value, and all mechanisms avail-able to the body for heat conservation andheat generation are maximally activatedto achieve this new “target value.” Forthis reason, although there are fewer sideeffects, maintaining normothermia canbe far more difficult (because of the


greater shivering response comparedwith the hypothermic state) than induc-ing hypothermia. Mayer et al (71) com-pared the efficacy of two surface coolingdevices to maintain normothermia; they

were able to maintain normothermia (de-fined as a core temperature �37.2°C) for59% of the treatment time in one groupand for 3% of the time in the other. Fever(temperature �38.3°C) occurred for 8%

of the time in one group and for 42% ofthe time in the other. The frequency andseverity of shivering was related to theefficacy of cooling (71). This underscoresthe difficulties that can be involved in

Table 5. Advantages, disadvantages, and time lag compared with gold standard of various temperature monitoring sites

SiteLevel ofAccuracy

Average Time Lag BetweenSite and Gold Standard Specific Advantages, Problems, and Limitations

Pulmonary artery High N/A �Highly precise and rapid temperature registration–Complex insertion procedure required–Needs to be removed after 72–96 hrs

Jugular bulb High N/A �Highly precise temperature registration�Venous blood coming directly from the brain reflects brain

temperature even more accurately than pulmonary arterycatheter measurements

–Complex insertion procedure required–Experimental form of neuromonitoring; used relatively rarely

Esophagus High 5 mins (range, 3–10) �Most rapid and accurate reflection of gold standard–Moderate risk of downward dislocation to stomach. This can

lead to a longer time lag and a slight (1°C–3°C) drop inregistered core temperature. As this deviation is relativelysmall, it may go unnoticed. Can be prevented by preciseinsertion to a depth of 32–38 cm

–Potential interference of diagnostic/therapeutic procedures suchas insertion of gastric feeding tubes, transesophagealechocardiography, gastroscopy, etc

�Occasionally problematic probe insertion procedureBladder Fair/high 20 mins (range, 10–60)a �Fairly easy probe insertion procedure �

�Low risk of dislocation�Combination with procedure (catheter insertion) that needs to

take place anyway–Relatively long time lag–Readings affected by rate of diuresis (which may be low in

some patients after cardiac arrest)–Sensor can move into saline-filled balloon at tip of some

catheters, affecting temperature readingsRectum Fair/high 15 mins (range, 10–40)b �Quick and easy probe insertion procedure

–High risk of dislocation (however, dislocation is likely to benoticed immediately because of the magnitude of thedifference with “true” core temperature)

–Relatively long time lagNasopharynx High 8 mins (range, 5–10) �Relatively quick and easy probe placement procedure

�Relatively short lag time�May reflect brain temperature better than more distant sites–Risk of probe misplacement �–Risk of nasal bleeding especially in patients receiving

anticoagulantsTympanic membrane Moderate/fair 10 mins (range, 5–20) �Very quick and easy probe placement procedure

–May reflect brain temperature better than more distant sites–Readings may be inaccurate�Continuous measurement may be uncomfortable for awake

patientsForehead Fair/highc 5 mins?c �Good correlation (r2 � .87) shown between temperature

measured by device and temp at 18 mm depth from skinsurface (i.e., in the frontal lobe of the brain)

–Good correlation with blood and esophageal temperature–Not yet well studied under conditions of rapid hypothermia

induction (i.e., rapid temperature change) or duringprolonged hypothermic conditions

–Requires �15 mins of calibration timePeripheral sites (axilla,

groin, etc.)Completely

inaccurateNo correlation with gold

standardPeripheral measurements should never be used to guide

hypothermia treatment. During hypothermia such readingsare completely inaccurate

aIn case of severe shock, oliguria, etc; bin case of severe shock; cseveral studies report good correlation of “deep forehead” temperature measured byCoretemp R device with blood and esophageal temperatures (183), but no studies have been performed during rapid induction of hypothermia or duringhypothermia maintenance (only in the perioperative setting).


maintaining normothermia. This appliesespecially to cooling awake patients,where it is often more difficult to controlshivering because high drug doses canlead to respiratory problems. Cooling canlead to a significant stress response if noanti-shivering treatment is given, or ifthe treatment is ineffective. Lenhardt etal (150) induced fever in nine healthyvolunteers by infusing interleukin-2, andsubsequently assessed the effects of sur-face cooling. They reported a 35% to 40%increase in oxygen consumption, an in-crease in catecholamine levels and signif-icant patient discomfort associated withcooling.

In the clinical setting such responsescan be counteracted with various anti-shivering drugs (Table 4). For the reasonsoutlined above, higher doses will usuallybe needed to control shivering duringinduction of normothermia than for hy-pothermia. This can be a problem espe-cially when cooling awake and nonventi-lated patients. Two recent studiesattempted to decrease myocardial injuryin patients with myocardial infarctionundergoing percutaneous coronary inter-vention, by inducing hypothermia beforereperfusion; in both of these studies amajority of patients failed to reach targettemperature before reperfusion. Interest-ingly, apparent benefits were observed inpatients who did reach target tempera-ture in these studies (1, 152).

Some studies have reported that shiv-ering thresholds can be reduced bywarming the hands, feet, and/or face ofthe patient, reducing the required dosesof antishivering drugs (60, 153–155).However, others have reported no or onlyminor effects of hand/face warming onshivering thresholds (99, 156). There issome in unanethetized humans evidencethat use of endovascular cooling methodsto maintain normothermia reduces theshivering response (157), although no di-rect comparative studies have been per-formed to address this issue.

Apart from the cooling devices listedin Table 2, hyperthermia can also betreated with various antipyretic drugssuch as acetaminophen. However, the ef-fectiveness of these drugs is limited, es-pecially in patients with neurologic (cen-tral) fever (158–164). In large studies theaverage decrease in temperature duringtreatment with high doses (4000–6000mg/day) of acetaminophen is about0.3°C–0.4°C (158–160). Similar resultshave been reported using high doses ofaspirin (163), and in small studies with

metamizol (161). Ibuprofen appears to beineffective in reducing core temperature(164).

Despite the relatively minor effects,administration of fever-suppressingdrugs is a helpful adjunctive treatmentbecause it induces a drop in temperaturewithout activating counter-regulatorymechanisms. The reasons for this and theprecise mechanisms of action of antipy-retics are beyond the scope of this review;this topic has been reviewed elsewhere(182). Thus, acetaminophen can be usedas an accessory method to reduce hyper-thermia in neurologic injury, but addi-tional cooling methods will be requiredto achieve normothermia in most pa-tients.

These observations lead to the follow-ing clinical scenario for the treatment ofcomatose survivors of witnessed cardiacarrest. Initiate infusion of cold (4°C) flu-ids using a pressure bag as soon as pos-sible, preferably in the emergency room.A cooling device should be connected tothe patient as soon as possible. The pres-ence of arrhythmias or cardiogenic shockshould not be regarded as counterindica-

tions for the use of mild hypothermia.percutaneous coronary interventionshould be performed as rapidly as possi-ble if acute myocardial infarction is sus-pected; cooling can be safely used in com-bination with it (168). A bolus dose ofmagnesium (2–3 g given over a period of5–10 minutes) can be given to preventshivering and cooling-associated hypo-magnesemia. A (low) maintenance doseof sedation should be given. Maintenancedoses of many drugs, especially sedatives,long-acting opiates, and muscle paralyz-ers, can and should be reduced in thisphase to avoid cumulation caused by hy-pothermia-related reductions in drugclearance rates. If shivering occurs thiscan be controlled with bolus doses ofmagnesium (2–3 g given over 5–10 min-utes) or fentanyl (50–100 g rapid bolusdose), or one of the other drugs listed inTable 4. Lab samples should be drawnevery 30 minutes in the induction phaseand every 4–6 hours in the maintenancephase, with special attention to electro-lyte and glucose levels and the blood gasanalysis. Antibiotic prophylaxis should beconsidered to avoid infections; blood cul-

Table 6. Practical checklist of issues to address and to avoid during therapeutic temperaturemanagement

Checklist for induced hypothermiaUse cooling device with central feedback loop to control temperature. Use core temperature

measurements to guide treatment. Add cold fluids in induction phase in most patients. Usepressure bag for administering cold fluids

Avoid hypovolemia and hypotension (cooling causes cold diuresis)Avoid electrolyte disorders (cooling causes loss of K, Mg, P; rapid rewarming can cause

hyperkalemia)Avoid hyperglycemia (cooling causes insulin resistance and decreased insulin secretion)Control shivering (options: magnesium; meperidine; quick-acting opiates; propofol;

benzodiazepines. Alternatives: clonidine; ketanserin; tramadol; urapidil; doxapram)Avoid skin damage/bedsores (prolonged direct exposure of the skin to ice or ice packs may cause

burns. Cooling causes vasoconstriction in the skin)Avoid/promptly treat infections (diagnosis can be difficult due to suppression of symptoms such

as fever and leukocytosis. Consider antibiotic prophylaxis, perform frequent cultures of bloodand other sites)

Use appropriate sedation and analgesia (animal data suggest loss of protective effects if sedationis insufficient; sedation also facilitates cooling by preventing shivering and causingvasodilation)

Adjust ventilator settings (cooling causes 2O2 consumption and 2CO2 production)Adjust feeding rate (cooling decreases metabolism by 7% to 10% per °C decrease below 37°C)Adjust drug dosage (drug clearance may change, including clearance of

sedatives/opiates/paralyzers; use bolus doses during hypothermia induction phase, avoid highmaintenance doses)

Don’t let core temperature fall below 30°C (risk of arrhythmias arises at temp �28°C–30°C)Don’t “overtreat” (bradycardia, mild metabolic acidosis, slight rise in lactate levels, liver enzymes

and amylase are normal consequences of hypothermia)Long-term paralysis is usually unnecessary (paralysis will mask inadequate sedation, and can

have adverse consequences such as increased risk of critical illness polyneuromyopathy).Paralysis can be considered in the induction phase to facilitate cooling and increase cooling speeds

Consider platelet administration before surgery or invasive procedures during coolingDon’t rewarm too quickly! (maximum 0.2°C–0.5°C/hr, slower in traumatic brain injury)Be aware that insulin requirements may decrease during rewarming (risk of hypoglycemia).General measures: provide good basic intensive care (many of the side effects of cooling can be

prevented or controlled with proper care)


tures should be drawn once a day duringhypothermia or controlled normother-mia treatment. After 24 hours slow re-warming (�0.25°C/hr) should be initi-ated. Hypoglycemia may develop in thisphase, and there is a risk for hyperkale-mia if rewarming is too rapid (�0.5°C/hr), and/or if the patient has gross renalfailure.

CONCLUSION

Induction of hypothermia has wide-ranging effects and will induce numerousphysiologic changes in virtually every or-gan in the body. Physicians and nursesapplying induced hypothermia in the in-tensive care unit need to be aware of thephysiologic and pathophysiologicalchanges and side effects associated withhypothermia, and should be familiar withthe available cooling techniques. A“checklist” of the most important precau-tions and countermeasures is provided inTable 6.

The success of hypothermia treat-ments will be determined to a large ex-tent by our ability to prevent or effec-tively deal with its side effects; thisapplies especially if hypothermia is usedfor prolonged periods. Devising compre-hensive protocols with detailed recom-mendations on rates of cooling and re-warming, as well as on the propermanagement of side effects, will be thekey to success. This will help us apply thishighly promising treatment more effec-tively, and perhaps to expand its usage toother areas such as traumatic brain in-jury and stroke.

REFERENCES

1. Polderman KH: Induced hypothermia andfever control for prevention and treatmentof neurological injuries. Lancet 2008; 371:1955–1969

2. Nolan JP, Deakin CD, Soar J, et al: Euro-pean Resuscitation Council guidelines forresuscitation 2005. Section 4. Adult ad-vanced life support. Resuscitation 2005;67(Suppl 1):S39–S86

3. Polderman KH: Application of therapeutichypothermia in the ICU: Opportunities andpitfalls of a promising treatment modality,Part 1. Indications and evidence. IntensiveCare Med 2004; 30:556–575

4. Bernard SA, Buist M: Induced hypothermiain critical care medicine: A review. CritCare Med 2003; 31:2041–2051

5. Zeiner A, Holzer M, Sterz F, et al: Hyper-thermia after cardiac arrest is associatedwith an unfavorable neurologic outcome.Arch Intern Med 2001; 161:2007–2012

6. Diringer MN, Reaven NL, Funk SE, et al:Elevated body temperature independentlycontributes to increased length of stay inneurologic intensive care unit patients. CritCare Med 2004; 32:1489–1495

7. Kammersgaard LP, Jorgensen HS, RungbyJA, et al: Admission body temperature pre-dicts long-term mortality after acute stroke:The Copenhagen stroke study. Stroke 2002;33:1759–1762

8. Oliveira-Filho J, Ezzeddine MA, Segal AZ, etal: Fever in subarachnoid hemorrhage: Re-lationship to vasospasm and outcome. Neu-rology 2001; 56:1299–1304

9. Commichau C, Scarmeas N, Mayer SA: Riskfactors for fever in the neurologic intensivecare unit. Neurology 2003; 60:837–841

10. Mayer SA, Sessler DI (Eds): Therapeutic Hy-pothermia. First Edition. New York NY,Marcel Dekker, 2005

11. Hayashi N, Dietrich DW (Eds): Brain Hypo-thermia Treatment. First Edition. Springer-Verlag, Tokyo, 2004

12. Polderman KH: Application of therapeutichypothermia in the intensive care unit. Op-portunities and pitfalls of a promising treat-ment modality, Part 2. Practical aspects andside effects. Intensive Care Med 2004; 30:757–769

13. Lopez M, Sessler DI, Walter K, et al: Rateand gender dependence of the sweating, va-soconstriction, and shivering thresholds inhumans. Anesthesiology 1994; 80:780–788

14. Sessler DI, Moayeri A, Stoen R, et al: Ther-moregulatory vasoconstriction decreasescutaneous heat loss. Anesthesiology 1990;73:656–660

15. Frank SM, Fleisher LA, Breslow MJ, et al:Perioperative maintenance of normother-mia reduces the incidence of morbid car-diac events. A randomized clinical trial.JAMA 1997; 277:1127–1134

16. Frank SM, Beattie C, Christopherson R, etal: Unintentional hypothermia is associatedwith postoperative myocardial ischemia.The Perioperative Ischemia RandomizedAnesthesia Trial Study Group. Anesthesiol-ogy 1993; 78:468–476

17. Leslie K, Sessler DI: Perioperative hypo-thermia in the high-risk surgical patient.Best Pract Res Clin Anaesthesiol 2003; 17:485–498

18. De Witte J, Sessler DI: Perioperative shiver-ing: Physiology and pharmacology. Anes-thesiology 2002; 96:467–484

19. Horvath SM, Spurr GB, Hutt BK, et al:Metabolic cost of shivering. J Appl Physiol1956; 8:595–602

20. Spurr GB, Hutt BK, Horvath SM: Shivering,oxygen consumption and body tempera-tures in acute exposure of men to two dif-ferent cold environments. J Appl Physiol1957; 11:58–64

21. Horvath SM, Radcliffe CE, Hutt BK, et al:Metabolic responses of old people to a coldenvironment. J Appl Physiol 1955; 8:145–148

22. Horvath SM, Hutt BK, Spurr GB, et al:Some metabolic responses of dogs having

low body temperature. Science 1953; 118:100–101

23. Frank SM, Fleisher LA, Olson KF, et al:Multivariate determinates of early postoper-ative oxygen consumption: The effects ofshivering, core temperature, and gender.Anesthesiology 1995; 83:241–249

24. Matsukawa T, Sessler DI, Sessler AM, et al:Heat flow and distribution during inductionof general anesthesia. Anesthesiology 1995;82:662–673

25. Hirsch LJ, Kull LL: Continuous EEG mon-itoring in the intensive care unit. Am JElectroneurodiagnostic Technol 2004; 44:137–158

26. Varelas PN, Mirski MA: Management of sei-zures in critically ill patients. Curr NeurolNeurosci Rep 2004; 4:489–496

27. Thoresen M, Satas S, Loberg EM, et al:Twenty-four hours of mild hypothermia inunsedated newborn pigs starting after a se-vere global hypoxic-ischemic insult is notneuroprotective. Pediatr Res 2001; 50:405–411

28. Tooley JR, Satas S, Porter H, et al: Headcooling with mild systemic hypothermia inanesthetized piglets is neuroprotective. AnnNeurol 2003; 53:65–72

29. Kammersgaard LP, Rasmussen BH, Jør-gensen HS, et al: Feasibility and safety ofinducing modest hypothermia in awake pa-tients with acute stroke through surfacecooling: A case-control study: The Copen-hagen Stroke Study. Stroke 2000; 31:2251–2256

30. Guluma KZ, Hemmen TM, Olsen SE, et al:A trial of therapeutic hypothermia via en-dovascular approach in awake patients withacute ischemic stroke: Methodology. AcadEmerg Med 2006; 13:820–827

31. Mahmud E, Keramati S: Highlights of the2003 transcatheter cardiovascular thera-peutics annual meeting: Clinical implica-tions. J Am Coll Cardiol 2004; 43:684–690

32. Matsukawa T, Kurz A, Sessler DI, et al:Propofol linearly reduces the vasoconstric-tion and shivering thresholds. Anesthesiol-ogy 1995; 82:1169–1180

33. Robinson BJ, Ebert TJ, O’Brien TJ, et al:Mechanisms whereby propofol mediates pe-ripheral vasodilation in humans: Sympa-thoinhibition or direct vascular relaxation?Anesthesiology 1997; 86:64–72

34. Toyota K, Sakura S, Saito Y, et al: The effectof pre-operative administration of midazolamon the development of intra-operative hypo-thermia. Anaesthesia 2004; 59:116–121

35. Matsukawa T, Hanagata K, Ozaki M, et al:I.m. midazolam as premedication producesa concentration-dependent decrease in coretemperature in male volunteers. Br J An-aesth 1997; 78:396–399

36. Ebert TJ, Ficke DJ, Arain SR, et al: Vasodi-lation from sufentanil in humans. AnesthAnalg 2005; 101:1677–1680

37. Gursoy S, Bagcivan I, Yildirim MK, et al:Vasorelaxant effect of opioid analgesics on


the isolated human radial artery. Eur J An-aesthesiol 2006; 23:496–500

38. Hsu HO, Hickey RF, Forbes AR: Morphinedecreases peripheral vascular resistance andincreases capacitance in man. Anesthesiol-ogy 1979; 50:98–102

39. Fay T: Early experiences with local and gen-eralized refrigeration of the human brain.J Neurosurg 1959; 16:239–259

40. Smith LW, Fay T: Observations on humanbeings with cancer maintained at reducedtemperatures of 75°–80° Fahrenheit. Am JClin Pathol 1940; 10:3–11

41. Dill DB, Forbes WH: Respiratory and meta-bolic effects of hypothermia. Am J Physiol1941; 132:685–697

42. Latronico N, Peli E, Botteri M: Critical ill-ness myopathy and neuropathy. Curr OpinCrit Care 2005; 11:126–132

43. Tortorici MA, Kochanek PM, Poloyac SM:Effects of hypothermia on drug disposition,metabolism, and response: A focus of hypo-thermia-mediated alterations on the cyto-chrome P450 enzyme system (review). CritCare Med 2007; 35:2196–2204

44. Sessler DI: Complications and treatment ofmild hypothermia. Anesthesiology 2001;95:531–543

45. Koren G, Barker C, Goresky G, et al: Theinfluence of hypothermia on the dispositionof fentanyl—Human and animal studies.Eur J Clin Pharmacol 1987; 32:373–376

46. Bansinath M, Turndorf H, Puig MM: Influ-ence of hypo and hyperthermia on disposi-tion of morphine. J Clin Pharmacol 1988;28:860–864

47. Leslie K, Sessler DI, Bjorksten AR, et al:Mild hypothermia alters propofol pharma-cokinetics and increases the duration of ac-tion of atracurium. Anesth Analg 1995; 80:1007–1014

48. Fukuoka N, Aibiki M, Tsukamoto T, et al:Biphasic concentration change during con-tinuous midazolam administration inbrain-injured patients undergoing thera-peutic moderate hypothermia. Resuscita-tion 2004; 60:225–230

49. Diefenbach C, Abel M, Buzello W: Greaterneuromuscular blocking potency of atra-curium during hypothermic than duringnormothermic cardiopulmonary bypass.Anesth Analg 1992; 75:675–678

50. Caldwell JE, Heier T, Wright PMC, et al:Temperature-dependent pharmacokineticsand pharmacodynamics of vecuronium. An-esthesiology 2000; 92:84–93

51. Beaufort AM, Wierda JM, Belopavlovic M, etal: The influence of hypothermia (surfacecooling) on the timecourse of action and onthe pharmacokinetics of rocuronium in hu-mans. Eur J Anaesthesiol Suppl 1995; 11:95–106

52. Cotten MD, Brown TG: Effects of pressoramines and ouabain on the heart and bloodpressure during hypothermia. J PharmacolExp Ther 1957; 121:319–329

53. Booth BP, Brien JF, Marks GS, et al: Theeffects of hypothermic and normothermic

cardiopulmonary bypass on glyceryl trini-trate activity. Anesth Analg 1994; 78:848–856

54. Cheng C, Matsukawa T, Sessler DI, et al:Increasing mean skin temperature linerlyreduces the core-temperature thresholdsfor vasoconstriction and shivering in hu-mans. Anesthesiology 1995; 82:1160–1168

55. Koren G, Chin TW: Hypothermia, alkalosis,and barbiturate clearance. J Pediatr 1983;102:643–644

56. Schaible DH, Cupit GC, Swedlow DB, et al:High-dose pentobarbital pharmacokineticsin hypothermic brain-injured children.J Pediatr 1982; 100:655–660

57. Zhou JX, Liu J: The effect of temperature onsolubility of volatile anesthetics in humantissues. Anesth Analg 2001; 93:234–238

58. Rink RA, Gray I, Ruckert RR, et al: Theeffect of hypothermia on morphine metab-olism in an isolated perfused liver. Anesthe-siology 1956; 17:377–384

59. Sessler DI: Perioperative heat balance (re-view). Anesthesiology 2000; 92:578–596

60. Badadja N, Strongilis E, Prescutti M, et al:Metabolic benefits of surface counter warm-ing during therapeutic temperature modu-lation. Crit Care Med 2009, in press

61. Polderman KH, Peerdeman SM, Girbes ARJ:Hypophosphatemia and hypomagnesemiainduced by cooling in patients with severehead injury. J Neurosurg 2001; 94:697–705

62. Polderman KH, Rijnsburger ER, PeerdemanSM, et al: Induction of hypothermia in pa-tients with various types of neurologic in-jury with use of large volumes of ice-coldintravenous fluid. Crit Care Med 2005; 33:2744–2751

63. Maxwell WL, Watson A, Queen R, et al:Slow, medium, or fast re-warming follow-ing post-traumatic hypothermia therapy?An ultrastructural perspective. J Neuro-trauma 2005; 22:873–884

64. Alam HB, Rhee P, Honma K, et al: Does therate of rewarming from profound hypother-mic arrest influence the outcome in a swinemodel of lethal hemorrhage? J Trauma2006; 60:134–146

65. Hildebrand F, van Griensven M, GiannoudisP, et al: Effects of hypothermia and re-warming on the inflammatory response in amurine multiple hit model of trauma.Cytokine 2005; 31:382–393

66. Kawahara F, Kadoi Y, Saito S, et al: Slowrewarming improves jugular venous oxygensaturation during rewarming. Acta Anaes-thesiol Scand 2003; 47:419–424

67. Lavinio A, Timofeev I, Nortje J, et al: Cere-brovascular reactivity during hypothermiaand rewarming. Br J Anaesth 2007; 99:237–244

68. Bissonnette B, Holtby HM, Davis AJ, et al:Cerebral hyperthermia in children after car-diopulmonary bypass. Anesthesiology 2000;93:611–618

69. Polderman KH, Girbes AJ: Central venouscatheter use. Part 1: Mechanical complica-tions. Intensive Care Med 2002; 28:1–17

70. Simosa HF, Petersen DJ, Agarwal SK, et al:Increased risk of deep venous thrombosiswith endovascular cooling in patients withtraumatic head injury. Am Surg 2007; 73:461–464

71. Mayer SA, Kowalski RG, Presciutti M, et al:Clinical trial of a novel surface cooling systemfor fever control in neurocritical care patients.Crit Care Med 2004; 32:2508–2515

72. Polderman KH: Keeping a cool head: Howto induce and maintain hypothermia. CritCare Med 2004; 32:2558–2560

73. Hoedemaekers CW, Ezzahti M, Gerritsen A,et al: Comparison of cooling methods toinduce and maintain normo- and hypother-mia in intensive care unit patients: A pro-spective intervention study. Crit Care 2007;11:R91

74. Polderman KH, Callaghan J. Equipment re-view: Cooling catheters to induce therapeu-tic hypothermia? Crit Care 2006; 10:234

75. Thoresen M, Whitelaw A: Cardiovascularchanges during mild therapeutic hypother-mia and rewarming in infants with hypoxic-ischemic encephalopathy. Pediatrics 2000;106:92–99

76. Reuler JB: Hypothermia: Pathophysiology,clinical settings, and management. Ann In-tern Med 1978; 89:519–527

77. Erecinska M, Thoresen M, Silver IA: Effectsof hypothermia on energy metabolism inMammalian central nervous system.J Cereb Blood Flow Metab 2003; 23:513–530

78. Hagerdal M, Harp J, Nilsson L, et al: Theeffect of induced hypothermia upon oxygenconsumption in the rat brain. J Neurochem1975; 24:311–316

79. Palmer C, Vannucci RC, Christensen MA, etal: Regional cerebral blood flow and glucoseutilization during hypothermia in newborndogs. Anesthesiology 1989; 71:730–737

80. Ehrlich MP, McCullough JN, Zhang N, et al:Effect of hypothermia on cerebral bloodflow and metabolism in the pig. Ann ThoracSurg 2002; 73:191–197

81. Aoki M, Nomura F, Stromski ME, et al:Effects of pH on brain energetics after hy-pothermic circulatory arrest. Ann ThoracSurg 1993; 55:1093–1103

82. Fischer UM, Cox CS Jr, Laine GA, et al: Mildhypothermia impairs left ventricular dia-stolic but not systolic function. J InvestSurg 2005; 18:291–296

83. Goldberg LI: Effects of hypothermia on con-tractility of the intact dog heart. Am JPhysiol 1958; 194:92–98

84. Suga H, Goto Y, Igarashi Y, et al: Cardiaccooling increases Emax without affectingrelation between O2 consumption and sys-tolic pressure-volume area in dog left ven-tricle. Circ Res 1988; 63:61–71

85. Mikane T, Araki J, Suzuki S, et al: O2 cost ofcontractility but not of mechanical energyincreases with temperature in canine leftventricle. Am J Physiol 1999; 277:H65–H73

86. Lewis ME, Al-Khalidi AH, Townend JN, etal: The effects of hypothermia on human


left ventricular contractile function duringcardiac surgery. J Am Coll Cardiol 2002;39:102–108

87. Moat NE, Lamb RK, Edwards JC, et al: In-duced hypothermia in the management ofrefractory low cardiac output states follow-ing cardiac surgery in infants and children.Eur J Cardiothorac Surg 1992; 6:579–584

88. Deakin CD, Knight H, Edwards JC, et al:Induced hypothermia in the postoperativemanagement of refractory cardiac failurefollowing paediatric cardiac surgery. Anaes-thesia 1998; 53:848–853

89. Dalrymple-Hay MJ, Deakin CD, Knight H,et al: Induced hypothermia as salvage treat-ment for refractory cardiac failure followingpaediatric cardiac surgery. Eur J Cardiotho-rac Surg 1999; 15:515–518

90. Moriyama Y, Iguro Y, Shimokawa S, et al:Successful application of hypothermia com-bined with intra-aortic balloon pump sup-port to low-cardiac-output state after openheart surgery. Angiology 1996; 47:595–599

91. Yahagi N, Kumon K, Watanabe Y, et al:Value of mild hypothermia in patients whohave severe circulatory insufficiency evenafter intra-aortic balloon pump. J ClinAnesth 1998; 10:120–125

92. Nessmann ME, Busch HM, Gundersen AL:Asystolic cardiac arrest in hypothermia. WisMed J 1983; 82:19–20

93. Kirby CK, Jensen JM, Johnson J: Defibrilla-tion of the ventricles under hypothermicconditions. AMA Arch Surg 1954; 68:663–665

94. Martinez JB, Kass I, Hoffman MS: Factorsinvolved in recovery of patient after pro-longed ventricular fibrillation during hypo-thermia. J Thorac Surg 1958; 36:749–756

95. Covino BG, Beavers WR: Changes in cardiaccontractility during immersion hypother-mia. Am J Physiol 1958; 195:433–436

96. Covino BG, Hegnauer AH: Hypothermicventricular fibrillation and its control. Sur-gery 1956; 40:475–480

97. Pozos RS, Danzl D: Human physiologicalresponses to cold stress and hypothermia.In: Medical Aspects of Harsh Environments,Vol Textbooks of Military Medicine. PandolfKB, Burr RE (Eds). Washington, DC, Bor-den Institute, Office of the Surgeon Gen-eral, US Army Medical Department 2001, pp351–382

98. Polderman KH, Tjong Tjin Joe R, Peerde-man SM, et al: Effects of artificially inducedhypothermia on intracranial pressure andoutcome in patients with severe traumatichead injury. Intensive Care Med 2002; 28:1563–1567

99. Van Zanten ARH, Polderman KH: Blowinghot and cold? Skin counterwarming to pre-vent shivering during therapeutic cooling.Crit Care Med 2009, in press

100. Nabel EG, Ganz P, Gordon JB, et al: Dilationof normal and constriction of atheroscle-rotic coronary arteries caused by the coldpressor test. Circulation 1988; 77:43–52

101. Hale SL, Kloner RA: Myocardial tempera-

ture in acute myocardial infarction: Protec-tion with mild regional hypothermia. Am JPhysiol 1997; 273(Part 2):H220–H227

102. Hale SL, Dae MW, Kloner RA: Hypothermiaduring reperfusion limits ‘no-reflow’ injuryin a rabbit model of acute myocardial in-farction. Cardiovasc Res 2003; 59:715–722

103. Hale SL, Kloner RA: Elevated body temper-ature during myocardial ischemia/reperfu-sion exacerbates necrosis and worsens no-reflow. Coron Artery Dis 2002; 13:177–181

104. Hale SL, Kloner RA: Myocardial tempera-ture reduction attenuates necrosis afterprolonged ischemia in rabbits. CardiovascRes 1998; 40:502–507

105. Hale SL, Dae MW, Kloner RA: Marked re-duction in no-reflow with late initiation ofhypothermia in a rabbit myocardial infarctmodel. J Am Coll Cardiol 2003; 41:381–382

106. Miki T, Liu GS, Cohen MV, et al: Mild hy-pothermia reduces infarct size in the beat-ing rabbit heart: A practical intervention foracute myocardial infarction? Basic Res Car-diol 1998; 93:372–383

107. Hale SL, Dave RH, Kloner RA: Regionalhypothermia reduces myocardial necrosiseven when instituted after the onset of isch-emia. Basic Res Cardiol 1997; 92:351–357

108. Dae MW, Gao DW, Sessler DI, et al: Effect ofendovascular cooling on myocardial tem-perature, infarct size, and cardiac output inhuman-sized pigs. Am J Physiol Heart CircPhysiol 2002; 282:H1584–H1591

109. Dixon SR, Whitbourn RJ, Dae MW, et al:Induction of mild systemic hypothermiawith endovascular cooling during primarypercutaneous coronary intervention foracute myocardial infarction. J Am Coll Car-diol 2002; 40:1928–1934

110. Kandzari DE, Chu A, Brodie BR, et al: Fea-sibility of endovascular cooling as an ad-junct to primary percutaneous coronary in-tervention (results of the LOWTEMP pilotstudy). Am J Cardiol 2004; 93:636–639

111. Sabiston DC Jr, Theilen EO, Gregg DE: Re-lationship of coronary blood flow and car-diac output and other parameters in hypo-thermia. Surgery 1955; 38:498–505

112. Jude JR, Haroutunian LM, Folse R: Hypo-thermic myocardial oxygenation. Am JPhysiol 1957; 190:57–62

113. McIntosh TK Vink R, Yamakami I, et al:Magnesium protects against neurologicaldeficit after brain injury. Brain Res 1989;482:252–260

114. Polderman KH, Zanten ARH van, GirbesARJ: The importance of magnesium in crit-ically ill patients: A role in mitigating neu-rological injury and in the prevention ofvasospasms. Intensive Care Med 2003; 29:1202–1203

115. Van den Bergh WM, Algra A, van Kooten F,et al: Magnesium sulfate in aneurysmal sub-arachnoid hemorrhage: A randomized con-trolled trial. Stroke 2005; 36:1011–1015

116. Altman D, Carroli G, Duley L, et al: Dowomen with pre-eclampsia, and their ba-bies, benefit from magnesium sulphate?

The Magpie Trial: A randomised placebo-controlled Trial. Lancet 2002; 359:1877–1890

117. Woods KL, Fletcher S, Roffe C: Intravenousmagnesium sulphate in suspected acutemyocardial infarction: Results of the secondLeichester Intravenous Magnesium Inter-vention Trial (LIMIT-II). Lancet 1992; 339:1553–1558

118. Abraham AS, Eylath U, Weinstein M, et al:Serum magnesium levels in patients withacute myocardial infarction. N Engl J Med1977; 296:862–863

119. Soliman HM, Mercan D, Lobo SS, et al:Development of ionized hypomagnesemia isassociated with higher mortality rates. CritCare Med 2003; 31:1082–1087

120. Rubeiz GJ, Thill-Baharozian M, Hardie D, etal: Association of hypomagnesaemia andmortality in acutely ill medical patients.Crit Care Med 1993; 21:203–209

121. Chernow B, Bamberger S, Stoiko M, et al:Hypomagnesaemia in patients in postoper-ative intensive care. Chest 1989; 95:391–397

122. Temkin NR, Anderson GD, Winn HR, et al:Magnesium sulfate for neuroprotection af-ter traumatic brain injury: A randomisedcontrolled trial. Lancet Neurol 2007;6:29–38

123. van den Berghe G, Wouters P, Weekers F, etal: Intensive insulin therapy in the criticallyill patients. N Engl J Med 2001; 345:1359–1367

124. Van den Berghe G, Wilmer A, Hermans G, etal: Intensive insulin therapy in the medicalICU. N Engl J Med 2006; 354:449–461

125. Auer RN: Non-pharmacologic (physiologic)neuroprotection in the treatment of brainischemia. Ann NY Acad Sci 2001; 939:271–282

126. Lundgren J, Smith ML, Siesjo BK: Influ-ence of moderate hypothermia on ischemicbrain damage incurred under hyperglyce-mic conditions. Exp Brain Res 1991; 84:91–101

127. De Courten-Myeres GM, Kleinholz M,Wagner KR, et al: Normoglycemia (not hy-poglycemia) optimizes outcome from mid-dle cerebral artery occlusion. J Cereb BloodFlow Metab 1994; 14:227–236

128. Capes SE, Hunt D, Malmberg K, et al: Stresshyperglycemia and prognosis of stroke innon-diabetic and diabetic patients: A sys-tematic overview. Stroke 2001; 32:2426–2432

129. Polderman KH, Girbes ARJ: Intensive insu-lin therapy: Of harm and health, of hypesand hypoglycemia. Crit Care Med 2006; 34:246–248

130. Brunkhorst FM, Engel C, Bloos F, et al:German Competence Network Sepsis (Sep-Net). Intensive insulin therapy and pen-tastarch resuscitation in severe sepsis.N Engl J Med 2008; 358:125–139

131. Wiener RS, Wiener DC, Larson RJ: Benefitsand risks of tight glucose control in criti-


cally ill adults: A meta-analysis. JAMA 2008;300:933–944

132. Current Controlled Trials: A multi-centre,open label, randomised controlled trial oftwo target ranges for glycaemic control inintensive care unit (ICU) patients. Availableat: http://www.controlled-trials.com/ISRCTN04968275/NICE-sugar

133. Bacher A: Effects of body temperature onblood gases. Intensive Care Med 2005; 31:24–27

134. Kern FH, Greeley WJ: Pro: pH-stat manage-ment of blood gases is not preferable toalpha-stat in patients undergoing braincooling for cardiac surgery. J CardiothoracVasc Anesth 1995; 9:215–218

135. Burrows FA: Con: pH-stat management ofblood gases is preferable to alpha-stat inpatients undergoing brain cooling for car-diac surgery. J Cardiothorac Vasc Anesth1995; 9:219–221

136. Laussen PC: Optimal blood gas manage-ment during deep hypothermic paediatriccardiac surgery: Alpha-stat is easy, but pH-stat may be preferable. Paediatr Anaesth2002; 12:199–204

137. Schaller B, Graf R: Hypothermia and stroke:The pathophysiological background. Patho-physiology 2003; 10:7–35

138. Kofstad J: Blood gases and hypothermia:Some theoretical and practical consider-ations. Scand J Clin Lab Invest Suppl 1996;224:21–26

139. Michelson AD, MacGregor H, Barnard MR,et al: Hypothermia-induced reversibleplatelet dysfunction. Thromb Haemost1994; 71:633–640

140. Watts DD, Trask A, Soeken K, et al: Hypo-thermic coagulopathy in trauma: Effect ofvarying levels of hypothermia on enzymespeed, platelet function, and fibrinolytic ac-tivity. J Trauma 1998; 44:846–854

141. Valeri CR, MacGregor H, Cassidy G, et al:Effects of temperature on bleeding time andclotting time in normal male and femalevolunteers. Crit Care Med 1995; 23:698–704

142. Patt A, McCroskey B, Moore E: Hypother-mia-induced coagulopathies in trauma (Re-view). Surg Clin North Am 1988; 68:775–785

143. Ferrara A, MacArthur JD, Wright HK, et al:Hypothermia and acidosis worsen coagu-lopathy in the patients requiring massivetransfusion. Am J Surg 1990; 160:515–518

144. Reed RL, Bracey AW, Hudson JD, et al:Hypothermia and blood coagulation: Disso-ciation between enzyme activity and clot-ting factor levels. Circ Shock 1990; 32:141–152

145. Valeri CR, Feingold H, Cassidy G, et al:Hypothermia-induced reversible plateletdysfunction. Ann Surg 1987; 205:175–181

146. Suehiro E, Fujisawa H, Akimura T, et al:Increased matrix metalloproteinase-9 inblood in association with activation of in-terleukin-6 after traumatic brain injury: In-

fluence of hypothermic therapy. J Neuro-trauma 2004; 21:1706–1711

147. de Jonge E, Schultz MJ, Spanjaard L, et al:Effects of selective decontamination of di-gestive tract on mortality and acquisition ofresistant bacteria in intensive care: A ran-domised controlled trial. Lancet 2003; 362:1011–1016

148. Polderman KH, Ely EW, Badr AE, et al:Induced hypothermia in traumatic brain in-jury: Considering the conflicting results ofmeta-analyses and moving forward. Inten-sive Care Med 2004; 30:1860–1864

149. Kurz A, Sessler DI, Lenhardt R, et al: Peri-operative normothermia to reduce the inci-dence of surgical-wound infection andshorten hospitalization. N Engl J Med 1996;334:1209–1215

150. Lenhardt R, Negishi C, Sessler DI, et al: Theeffects of physical treatment on induced fe-ver in humans. Am J Med 1999; 106:550–555

151. Akata T, Setoguchi H, Shirozu K, et al:Reliability of temperatures measured atstandard monitoring sites as an index ofbrain temperature during deep hypother-mic cardiopulmonary bypass conducted forthoracic aortic reconstruction. J ThoracCardiovasc Surg 2007; 133:1559–1565

152. COOL MI II: Cooling as an adjunctive ther-apy to percutaneous intervention in pa-tients with acute myocardial infarction.ClinicalTrials.gov Identifier: NCT00248196.Available at: http://clinicaltrials.gov/ct/search?term�COOL-MI&submit�Search

153. Sweney MT, Sigg DC, Tahvildari S, et al:Shiver suppression using focal hand warm-ing in unanesthetized normal subjects. An-esthesiology 2001; 95:1089–1095

154. Kimberger O, Ali SZ, Markstaller M, et al:Meperidine and skin surface warming addi-tively reduce the shivering threshold: A vol-unteer study. Crit Care 2007; 11:R29

155. Laizzo PA, Jeon YM, Sigg DC: Facial warm-ing increases the threshold for shivering.J Neurosurg Anesthesiol 1999; 11:231–239

156. Doufas AG, Wadhwa A, Lin CM, et al: Nei-ther arm nor face warming reduces theshivering threshold in unanesthetized hu-mans. Stroke 2003; 34:1736–1740

157. Schmutzhard E, Engelhardt K, Beer R, etal: Safety and efficacy of a novel intravascu-lar cooling device to control body tempera-ture in neurologic intensive care patients: Aprospective pilot study. Crit Care Med 2002;30:2481–2488

158. Dippel DW, van Breda EJ, van Gemert HM,et al: Effect of paracetamol (acetamino-phen) on body temperature in acute isch-emic stroke: A double-blind, randomizedphase II clinical trial. Stroke 2001; 32:1607–1612

159. Dippel DW, van Breda EJ, van der Worp HB,et al: Effect of paracetamol (acetamino-phen) and ibuprofen on body temperature

in acute ischemic stroke PISA, a phase IIdouble-blind, randomized, placebo-con-trolled trial �ISRCTN98608690�. BMC Car-diovasc Disord 2003; 3:2

160. Kasner SE, Wein T, Piriyawat P, et al: Acet-aminophen for altering body temperaturein acute stroke: A randomized clinical trial.Stroke 2002; 33:130–134

161. Gozzoli V, Treggiari MM, Kleger GR, et al:Randomized trial of the effect of antipyresisby metamizol, propacetamol or externalcooling on metabolism, hemodynamics andinflammatory response. Intensive Care Med2004; 30:401–407

162. Henker R, Rogers S, Kramer DJ, et al: Com-parison of fever treatments in the criticallyill: A pilot study. Am J Crit Care 2001;10:276–280

163. Bachert C, Chuchalin AG, Eisebitt R, et al:Aspirin compared with acetaminophen inthe treatment of fever and other symptomsof upper respiratory tract infection inadults: A multicenter, randomized, double-blind, double-dummy, placebo-controlled,parallel-group, single-dose, 6-hour dose-ranging study. Clin Ther 2005; 27:993–1003

164. Walson PD, Galletta G, Braden NJ, et al:Ibuprofen, acetaminophen, and placebotreatment of febrile children. Clin Pharma-col Ther 1989; 46:9–17

165. Ying CL, Tsang SF, Ng KF: The potentialuse of desmopressin to correct hypother-mia-induced impairment of primary hae-mostasis—An in vitro study using PFA-100.Resuscitation 2008; 76:129–133

166. Hovdenes J, Laake JH, Aaberge L, et al:Therapeutic hypothermia after out-of-hospital cardiac arrest: Experiences withpatients treated with percutaneous coro-nary intervention and cardiogenic shock.Acta Anaesthesiol Scand 2007; 51:137–142

167. Skulec R, Kovarnik T, Dostalova G, et al:Induction of mild hypothermia in cardiacarrest survivors presenting with cardio-genic shock syndrome. Acta AnaesthesiolScand 2008; 52:188–194

168. Wolfrum S, Pierau C, Radke PW, et al: Mildtherapeutic hypothermia in patients afterout-of-hospital cardiac arrest due to acuteST-segment elevation myocardial infarctionundergoing immediate percutaneous coro-nary intervention. Crit Care Med 2008; 36:1780–1786

169. Frank SM, Satitpunwaycha P, Bruce SR, etal: Increased myocardial perfusion and sym-pathoadrenal activation during mild corehypothermia in awake humans. Clin Sci(Lond) 2003; 104:503–508

170. Frank SM, Higgins MS, Breslow MJ, et al:The catecholamine, cortisol, and hemody-namic responses to mild perioperative hy-pothermia. A randomized clinical trial. An-esthesiology 1995; 82:83–93

171. Chi OZ, Choi YK, Lee DI, et al: Intraopera-tive mild hypothermia does not increase theplasma concentration of stress hormones


during neurosurgery. Can J Anaesth 2001;48:815–818

172. Firmin RK, Bouloux P, Allen P, et al: Sym-pathoadrenal function during cardiac oper-ations in infants with the technique of sur-face cooling, limited cardiopulmonarybypass, and circulatory arrest. J ThoracCardiovasc Surg 1985; 90:729–735

173. Wood M, Shand DG, Wood AJ: The sympa-thetic response to profound hypothermiaand circulatory arrest in infants. Can An-aesth Soc J 1980; 27:125–131

174. Kawada T, Kitagawa H, Yamazaki T, et al:Hypothermia reduces ischemia- and stimu-lation-induced myocardial interstitial nor-epinephrine and acetylcholine releases.J Appl Physiol 2007; 102:622–627

175. Boddicker KA, Zhang Y, Zimmerman MB, etal: Hypothermia improves defibrillationsuccess and resuscitation outcomes from

ventricular fibrillation. Circulation 2005;111:3195–3201

176. Harada M, Honjo H, Yamazaki M, et al:Moderate hypothermia increases the chanceof spiral wave collision in favor of self-termination of ventricular tachycardia/fibrillation. Am J Physiol Heart Circ Physiol2008; 294:H1896–H1905

177. Rhee BJ, Zhang Y, Boddicker KA, et al:Effect of hypothermia on transthoracic de-fibrillation in a swine model. Resuscitation2005; 65:79–85

178. Bash SE, Shah JJ, Albers WH, et al: Hy-pothermia for the treatment of postsurgi-cal greatly accelerated junctional ectopictachycardia. J Am Coll Cardiol 1987; 10:1095–1099

179. Balaji S, Sullivan J, Deanfield J, et al:Moderate hypothermia in the manage-ment of resistant automatic tachycar-

dias in children. Br Heart J 1991; 66:221–224

180. Pfammatter JP, Paul T, Ziemer G, et al:Successful management of junctionaltachycardia by hypothermia after cardiacoperations in infants. Ann Thorac Surg1995; 60:556–560

181. Mosquera Perez I, Rueda Nunez F, MedranoLopez C, et al: Management by hypothermiaof junctional ectopic tachycardia appearingafter pediatric heart surgery. Rev Esp Car-diol 2003; 56:510–514

182. Aronoff DM, Neilson EG: Antipyretics:Mechanisms of action and clinical use infever suppression. Am J Med 2001; 111:304–315

183. Harioka T, Matsukawa T, Ozaki M, et al:“Deep-forehead” temperature correlateswell with blood temperature. Can J Anaesth2000; 47:980–983


Therapeutic hypothermia and controlled normothermia in the ... · Therapeutic hypothermia and...

Documents

Transcript of Therapeutic hypothermia and controlled normothermia in the ... · Therapeutic hypothermia and...