n

SAIgGor

SD

J

Intravascular Cooling in the Treatment of Stroke (ICTuS):Early Clinical Experience

Patrick D. Lyden, MD,*,† Robin L. Allgren, MD, PhD,‡ Ken Ng, MD,§Paul Akins, MD,� Brett Meyer, MD,* Fahmi Al-Sanani, MD,¶ Helmi Lutsep, MD,#

John Dobak, MD,‡ Bradley S. Matsubara, MD,‡ and Justin Zivin, MD, PhD*,†

We sought to evaluate the safety and feasibility of mild therapeutic hypothermiausing an endovascular temperature management system in awake acute ischemicstroke patients. The Intravascular Cooling in the Treatment of Stroke (ICTuS) studywas an uncontrolled, multicenter development and feasibility study of consciouspatients (n � 18) presenting within 12 hours of onset of an acute ischemic stroke at5 clinical sites in the United States. Enrolled patients were to undergo coretemperature management using an endovascular cooling system to induce andmaintain mild, therapeutic hypothermia (target temperature of 33.0°C) for a periodof either 12 or 24 hours, followed by controlled rewarming to 36.5°C over thesubsequent 12-hour period. Nine patients underwent 12 hours of cooling followedby 12 hours of controlled rewarming, and 6 patients underwent 24 hours of coolingfollowed by 12 hours of controlled rewarming. Three patients underwent �1.5hours of hypothermia due to clinical or technical issues. We also developed anantishivering regimen using buspirone and meperidine administered prophylacti-cally to suppress shivering. The endovascular cooling catheter was well tolerated,with acceptable adverse event rates. Increasing the duration of hypothermiaadministration from 12 hours to 24 hours did not appear to increase the incidenceor severity of adverse effects. Endovascular cooling with a proactive antishiveringregimen can be accomplished in awake stroke patients. Further studies are neededto establish the safety and efficacy of this approach. Key Words: Hypothermia—stroke—clinical trial—thrombolysis.© 2005 by National Stroke Association

ctabnpes

pctbth

Hypothermia has proven to be a potent putativeeuroprotectant in preclinical experimental studies of

From the *Department of Neurosciences, University of Californiaan Diego School of Medicine, †Department of Neurology, Veteransdministration Medical Center, San Diego, ‡Innercool Therapies

nc., San Diego, California; §Department of Neurology, Ocala Re-ional Medical Center, Florida; �Department of Neurology, Mercyeneral Hospital, Sacramento, California; ¶Department of Neurol-gy, University of Texas, Houston, Texas; and #Department of Neu-ology, Oregon Health Sciences University, Portland, Oregon.

Received November 17, 2004; accepted December 28, 2004.Address reprint requests to Patrick D. Lyden, MD, FAAN, UCSD

troke Center, OPC Third Floor, Suite #3, 200 W. Arbor Drive, Saniego, CA 92103. E-mail: plyden@ucsd.edu.1052-3057/$—see front matter© 2005 by National Stroke Association

mdoi:10.1016/j.jstrokecerebrovasdis.2005.01.001

ournal of Stroke and Cerebrovascular Diseases, Vol. 14, No. 3 (May-Ju

erebral ischemia. Two clinical studies of patientsreated with 12-24 hours of hypothermia after cardiacrrest (a period of global cerebral ischemia) exhibitedetter neurologic outcomes and lower mortality thanormothermic controls.1,2 Thus hypothermia may holdromise as a treatment for acute ischemic stroke. How-ver, to be practical in the clinical setting, hypothermiahould be safe and easy to administer.

Previous clinical studies of hypothermia in strokeatients used surface cooling methods, such as surfaceooling blankets, ice packs, and alcohol baths.3-6 Al-hough these methods can cool patients, they are cum-ersome and slow, often requiring 3-8 hours to achievearget temperatures (although recent innovations mayave reduced the time to reach target to about 90

inutes).7 In addition, maintaining control at a desired

ne), 2005: pp 107-114 107

ttbtpma

inmdnveasd3

mslsasc

(m1T(tegFripopibtcat

ws

ssessm

P.D. LYDEN ET AL.108

emperature is difficult. For example, in 1 study, thearget temperature of 33°C was reached after 3.5 hours,ut in 9 of 10 patients there was an overshoot to cooleremperatures (as low as 28.4°C).4 Because lower tem-eratures can exacerbate adverse effects of hypother-ia, the ability to control patient temperatures and

void overcooling during hypothermia is important.Suppressing shivering during hypothermia is also crit-

cal, because shivering generates heat and can essentiallyegate the desired cooling. In most previous hypother-ia studies,4-6,8 to suppress shivering, patients were se-

ated (with, e.g., fenantyl and propofol), paralyzed (i.e.,euromuscular blockade), intubated, and mechanicallyentilated. These measures may be less well tolerated inlderly stroke patients than in relatively younger cardiacrrest patients. For example, in a study of surface cooling,hivering was controlled with intravenous (IV) meperi-ine in awake patients, but patients were cooled to only5.5°C over 5 hours.3

Developing better methods of administering hypother-ia in a controlled manner while adequately suppressing

hivering is an important prerequisite to the execution ofarge pivotal efficacy studies of hypothermia therapy introke. In the present study, we sought to develop a safend feasible method for inducing hypothermia in awaketroke patients. A similar experience using a different

Table 1. ICTuS inclu

Inclusion criteria1. Age 18 to 85 inclusive2. Symptoms of acute ischemic stroke still present by3. Stroke onset within 12 hours; awoke with stroke al

12 hours4. Any subtype of ischemic stroke with NIHSS �4 a

Exclusion criteria1. Etiology other than ischemic stroke2. Item 1a on NIHSS �13. Symptoms resolving or so mild that baseline NIHS4. Contraindications to hypothermia, such as patients

(e.g., cryoglobulinemia, Sickle cell disease, serum cthromboangiitis obliterans

5. Comorbid conditions likely to complicate therapy,a. End-stage cardiomyopathyb. Uncompensated arrhythmiac. Myopathyd. Liver disease severe enough to elevate bilirubine. History of pelvic or abdominal mass likely to cof. Dementia severe enough to prevent valid conseng. End-stage AIDS

6. Intracerebral hematoma; minimal hemorrhagic tran7. Any intraventricular hemorrhage8. Systolic blood pressure �210 or �100, diastolic b9. Severe coagulopathy, e.g., INR � 3.0 � control or

10. Pregnancy in women of childbearing potential, con11. Medical conditions likely to interfere with patient a

ooling device was published recently.9 m

Methods

The Intravascular Cooling in the Treatment of StrokeICTuS) study was a phase 1 prospective, uncontrolled,

ulticenter development/feasibility and safety study of2 or 24 hours of hypothermia in awake stroke patients.he planned sample size was �20 patients in 5 centers

see Appendix 1 for a list of the study sites). The studyarget population was awake patients with acute isch-mic stroke; the full inclusion and exclusion criteria areiven in Table 1. The treatment protocol is illustrated inigure 1. Patients meeting the inclusion/exclusion crite-ia were enrolled into the study after providing writtennformed consent, undergoing a complete history andhysical examination, and exhibiting acceptable resultsf the following tests: vital signs, oxygen saturation,rothrombin time, activated partial thromboplastin time,

nternational normalized ratio, serum electrolytes, biliru-in, blood urea nitrogen, creatinine, basic liver functionests, complete blood count with platelet count, brainomputed tomography (CT) or magnetic resonance im-ging scan, and neurologic examination using the Na-ional Institutes of Health Stroke Scale (NIHSS).10

Cooling to a target temperature of 33.0°C was initiatedithin 12 hours of stroke onset (patients awakening with

troke were timed from the last normal observation) and

nd exclusion criteria

e consent obtainedif time since last known to be symptom-free is within

me hypothermia begins

at the time hypothermia beginsnown hematologic dyscrasias that affect thrombosisglutinins) or vasospastic disorders such as Raynaud’s or

mple:

inferior vena cava

tion acceptable

essure �110 or �50 mmHg� 50 secondsby positive urine pregnancy testent

sion a

the timlowed

t the ti

S �4with kold ag

for exa

levelmpresst

sforma

lood prPTT

firmed

aintained for 12 or 24 hours. The 12-hour group was

c2m3tasdsch

Ccclhcpcarno

ostwstco

ptpeTtEctprattaor

oaamtseawpl(wtnopi

iopwdtut

FnA

INTRAVASCULAR COOLING IN THE TREATMENT OF STROKE 109

ompleted first, and then patients were enrolled into the4-hour cohort. At the conclusion of hypothermia treat-ent, patients were gradually rewarmed over 12 hours to

6.5°C (normothermia). Each patient was examined usinghe NIHSS, Barthel index (BI), and modified Rankin scalet the conclusion of treatment, 24 hours after the conclu-ion of treatment, on discharge from the hospital, and 30ays after stroke.11,12 A lower extremity venous ultra-ound examination was performed within 24 hours ofatheter removal. Brain CT scans were obtained 36 to 48ours after stroke onset and 30 days after the stroke.Endovascular cooling was performed with the Celsius

ontrol™ system (Innercool Therapies, San Diego, CA),onsisting of a catheter (#9 Fr) connected to a controlonsole. The catheter has a heparin-coated flexible metal-ic heat-transfer surface, which facilitates the extraction ofeat from the blood as it flows by the catheter. Theatheter is connected to a console that controls the tem-erature and flow rate of the fluid circulated within theatheter and adjusts them in response to patient temper-ture feedback (provided by an esophageal or nasopha-yngeal temperature probe connected to the console) aseeded to achieve and maintain the target hypothermic

igure 1. Study protocol: Outline of the basic procedures in ICTuS. N/A,ot applicable; mNIHSS, the abbreviated NIHSS used by the bedside nurse;E, adverse event; SAE, serious adverse event.

r normothermic temperature. t

A vascular access sheath was inserted into a femoral veinf each patient. The catheter was then inserted through theheath and advanced until the tip of the catheter rested inhe inferior vena cava, below the diaphragm. In patients

ho received tissue plasminogen activator (rt-PA) fortroke, sheath placement was delayed until 30 minutes afterhe rt-PA infusion ended. Proper catheter placement wasonfirmed with an abdominal flat plate radiograph or flu-roscopy before cooling.Two temperature-monitoring devices were placed, de-

ending on patient comfort, using 2 of 3 possible sites:ympanic membrane, bladder, and esophagus. The pur-ose was to determine the feasibility of the bladder orspophageal probes for feedback control of the console.he tympanic probe was planned as a safety backup in

he event that the controlling probe became disconnected.ach patient was monitored continuously in an intensiveare setting and checked hourly for hematoma or infec-ion at the catheter entry site, peripheral pulses, bloodressure, pulse, respiration, oxygen saturation, cardiachythm, and manual tympanic temperature. An abbrevi-ted NIHSS was obtained every 4 hours. Any deteriora-ion (e.g., a 2-point increase in the NIHSS) was reportedo the responsible investigator immediately. Similarly,ny potential adverse event, including vital signs outsidef normal parameters, was reported immediately to theesponsible investigator.

After catheter placement, a warming blanket (electricr convective air) was placed over the patient shortlyfter cooling began to reduce shivering. In addition, wettempted to design an antishivering protocol using IVeperidine. A conservative regimen was attempted in

he first 9 patients, but proved inadequate to controlhivering. The conservative regimen consisted of smallscalating doses of IV meperidine administered on ans-needed basis in response to shivering. The regimenas then modified for the subsequent patients to a moreroactive approach. The proactive protocol used prophy-

actic administration of loading doses of meperidinerange, 85-188 mg total loading dose, depending on

eight and gender) plus 30 mg of oral buspirone beforehe start of cooling, followed by a prophylactic mainte-ance infusion of meperidine (16-44 mg/hr) and 15 mg ofral buspirone every 8 hours throughout the coolingeriod.13 The combination of buspirone and meperidine

s thought to synergistically suppress shivering.13

All patients received antithrombotic prophylaxis withntermittent lower extremity compression, a combinationf oral aspirin and subcutaneous heparin, or both com-ression and aspirin/heparin. The compression methodas preferred. Unproven stroke therapies, such as high-ose steroids, mannitol, hyperventilation, or craniec-

omy, were not permitted. If thrombolytic therapy wassed, then the patient was required to be fully eligible for

hrombolysis according to the NINDS thrombolytic pro-

ocol.14 In particular, antiplatelet and anticoagulant drugs

wct

wssmhgdpvvhocs

r[1wit5a

fwcowa

coat(th

ms(c(awTp3i(t(

tt3fic–ch�

ta03ttv

Fpbod

P.D. LYDEN ET AL.110

ere prohibited for 24 hours after thrombolysis; onlyompression devices were allowed for deep venoushrombosis (DVT) prevention in these patients.

The primary purpose of this trial was to establishhether endovascular cooling was feasible in awake

troke patients. Important in this assessment was patientafety. Any adverse events occurring during hypother-ia, during rewarming, and after catheter removal until

ospital discharge were noted. Particular attention wasiven to venous clotting, puncture site hematoma, car-iac dysrythmias, and infectious and hematologic com-lications from the indwelling catheter. Any serious ad-erse events were noted through the 30-day follow-upisit. Follow-up CT brain scans were obtained 36-48ours after treatment and checked for evidence of hem-rrhage. All CT scans were forwarded to the coordinatingenter and read by 1 examiner blinded to the patients’tatus and treatment phase.

Results

A total of 18 patients (5 women and 13 men) wereecruited into the trial. The mean (� standard deviationSD]) patient age at the time of enrollment was 66.2 �

1.7 years (range, 38.9-81.4 years). Mean patient weightas 90.7 � 21.8 kg (range, 63-154 kg). The mean time

nterval from the onset of stroke symptoms to presenta-ion was 3.3 � 3.3 hours (range, 0.7-13.0 hours). Overall,patients (28%) presented within 3 hours of stroke onset

nd received rt-PA.All 18 enrolled patients underwent placement of a

emoral venous introducer sheath. Nine patients under-ent 12 hours of hypothermia followed by 12 hours of

ontrolled rewarming. Six patients underwent 24 hoursf hypothermia followed by 12 hours of controlled re-arming. In 1 patient, hypothermia was discontinued

igure 2. Cooling curves for patients treated with conservative versusroactive antishivering. Temperature #1: first recorded vitals; #2: coolingegins; #3: 1 hour into cooling; #4: lowest tolerated temperature; #5: startf rewarming; #6: end of rewarming. The groups are statistically significantlyifferent at the 1-hour and lowest-tolerated temperatures (#3 and #4).

fter 67 minutes due to the family’s withdrawal of study

onsent. Two patients underwent placement of the fem-ral introducer sheath for the catheter but did not receiveny hypothermia (due to hypotension in 1 and inabilityo correctly insert the catheter in the other). The mean� SD) time interval from the onset of stroke symptomso catheter introducer sheath insertion was 7.4 � 2.9ours (range, 3.1-15.0 hours).For the 16 patients for whom cooling was initiated, theean (� SD) time interval from the onset of stroke

ymptoms to the start of cooling was 8.0 �mn; 3.8 hoursrange, 4.3-20.3 hours). Of the 15 patients who wereooled according to protocol (i.e., for 12 or 24 hours), 853%) were treated with the initial, more conservativentishivering regimen; the remaining 7 patients (47%)ere treated with the proactive antishivering regimen.he mean (� SD) temperature on admission for all 15atients who were cooled was 36.5°C � 0.4°C (range,5.8-37.0°C). The average temperature when cooling wasnitiated for the conservative group was 36.7 � 1.0°Crange, 35.1-37.6°C). For patients in the proactive group,emperature at the start of cooling was 35.8 � 0.7°Crange, 34.8-36.8°C) (P � .41).

One hour into the cooling process, the mean � SDemperature was 36.1 � 0.7°C (range, 35.1�mn;37.0°C) inhe conservative group and 34.4 � 0.6°C (range, 33.3-5.5°C) in the proactive group (P � .00028). During therst hour of cooling, the proactive group had a signifi-antly faster average rate of temperature change of1.41 � 0.82°C/hr (range, �0.60°C/hr to �2.60° C/hr),ompared with just �0.56 � 0.58°C/hr (range, �0.30°C/r– �1.33°C/hr) for patients in the conservative group (P

.036). Similarly, the mean lowest achievable tempera-ure was significantly colder for the patients in the pro-ctive group than those in the conservative group: 33.7 �

.7°C (range, 32.7-34.6°C) versus 35.6 � 1.0°C (range,3.5-36.7°C) (P � .0012), respectively. However, the timeo reach the coldest temperature did not differ betweenhe cooling regimens, averaging 7 hours. Temperatureersus time in the patient groups treated with the con-

Table 2. All adverse events

Adverse event n

DVT 4Bradycardia 4Hypokalemia 4Nausea and/or vomiting 3Groin site hematoma 2Atrial fibrillation 2Catheter site pain 1Shivering 1Atrioventricular block 1Supraventricular tachycardia 1

Renal insufficiency 1

sg

rpasDseeitw3o1shi

beppa

afhttbtTtipD(p

INTRAVASCULAR COOLING IN THE TREATMENT OF STROKE 111

ervative and proactive antishivering regimens is shownraphically in Figure 2.Three deaths occurred during the study, none directly

elated to the catheter system, hypothermia, or studyrocedures (Table 2). No unanticipated device-relateddverse effects were reported. There were 11 reportederious adverse events in 9 patients, 1 of whom (a femoralVT) was considered related to the Celsius Control™

ystem or to hypothermia (Table 2). A total of 24 adversevents were reported (Table 3), 7 of which were consid-red possibly related to the catheter system: DVT (n � 4),nsertion site hematoma (n � 2), and pain at the cathe-erization site (n � 1). An additional 7 adverse events

ere considered hypothermia-related: bradycardia (n �

), nausea/vomiting due to IV meperidine given as partf the antishivering regimen (n � 3), and shivering (n �

). The 3 cases of bradycardia did not result in hypoten-ion. Prolonged administration of hypothermia (�1.5ours vs 12 hours vs 24 hours) was not associated with an

ncrease in bleeding, infectious, hemodynamic, or meta-

Table 4. All reporte

PatientID

Cooling(hrs)

Devicerelated?

Hypothermiarelated?* t-PA?

03-03 0 No No YesYes No Yes

05-02 1 No No No02-01 12 No No Yes03-01 12 No No No04-01 12 No No No

12 Yes No No04-02 12 No No No05-03 12 No No No05-07 24 No No Yes

SAE, serious adverse event.

Table 3. Ser

PatientID

Cooling(hrs)

Devicerelated?

Duringhypothermia?

03-03 0 No No05-02 1 No No01-01 12 Yes No04-01 12 No Yes05-03 12 No No01-02 24 No No

No No05-04 24 No Yes

No No05-06 24 No No05-07 24 No No

SAE, serious adverse event.

*No � “not related” or “probably not related”; Yes � “probably relat

olic complications. The incidence rates of these adverseffects reported here are in keeping with previouslyublished pilot studies of mild-to-moderate induced hy-othermia for the treatment of ischemic stroke or cardiacrrest.1-6,8

The 4 study patients with DVT were asymptomatic,nd the DVT was detected on the protocol-specifiedollow-up lower extremity ultrasound performed at 24ours. They all resolved without clinical sequelae after

reatment with heparin/coumadin therapy. Three ofhese DVTs were rated by the investigators as “proba-ly” or “definitely” related to the catheter use, becausehey occurred at the femoral site of catheter insertion.he fourth DVT was probably unrelated to the use of

he catheter, because it was found in the lower extrem-ty contralateral to the side in which the catheter waslaced. However, if one assumes that all 4 reportedVTs were catheter-related, the overall incidence

22%; 4/18) in this study is consistent with rates re-orted in the literature for femoral catheters, as well as

ding complications

E? Description

es Hemorrhagic transformation of infarct (fatal)o Hematoma, left groino GI bleeding (unspecified source)o GI bleeding (hematemesis, superficial gastric ulcer)o Hematuria (Foley catheter trauma)o Hemorrhagic transformation of infarcto Hematoma, right groino Hematuria (Foley catheter trauma)o Hematuriaes Hemorrhagic transformation of infarct (fatal)

verse events

day ofAE SAE

2 Hemorrhagic transformation of infarct (fatal)2 Neurologic decompensation (fatal)1 Femoral DVT0 Bradycardia

13 Pneumonia, aspiration14 Myocardial infarction17 Ventricular tachycardia

0 Absence seizure-like activity12 Hypertension, uncontrolled

6 Pulmonary embolism1 Hemorrhagic transformation of infarct (fatal)

d blee

SA

YNNNNNNNNY

ious ad

StudyS

ed” or “definitely related.”

Ds

(a(taatcth1Awbct

Hfib

aargcgn

vgbsrtg

gh42d1242trm

dhptta

Fthw70

P.D. LYDEN ET AL.112

VT rates reported for patients hospitalized for acutetroke.15,16

In the 1 patient who received hypothermia for �1 houra 73-year-old man with a known history of coronaryrtery disease), intermittent supraventricular tachycardiaSVT) was reported as an adverse event beginning closeo the initiation of cooling. This patient became agitatedfter administration of fentanyl and midazolam in prep-ration for catheter placement, which prompted his wifeo withdraw study consent after only 67 minutes ofooling and may have contributed to his SVT. Two pa-ients had the onset of atrial fibrillation (AF) duringypothermia. Both of these patients were enrolled in the2-hour group and had a known history of intermittentF; the arrhythmia required no treatment and resolvedithout sequelae in both cases. There were 2 cases of

radycardia in the 24-hour group that were present beforeooling started but increased in severity during the hypo-hermic period. Hypokalemia was reported in 4 patients.

All bleeding complications are summarized in Table 4.emorrhagic transformation of the incident cerebral in-

arction occurred in 3 of 18 patients (16.7%), as illustratedn Figure 3. Two patients died as a result of intracranialleeding; 1 patient was not cooled and did receive rt-PA,

Table 5. Infections

Reported adverse event

�1.5 hrs (n �

No.

Fever 1Leukocytosis 0Pneumonia, aspiration 1Upper respiratory tract infection 0Urinary tract infection 0

igure 3. Day-2 head CT scans for the 3 patients with hemorrhagicransformations. (A) A 72-year-old patient who received rt-PA within 3ours of stroke and no cooling (patient 03-03). (B) A 63-year-old patientho received no rt-PA and 12 hours of cooling (patient 04-01). (C) A4-year-old patient who received rt-PA and 24 hours of cooling (patient5-07).

Yeast infection 0 0%

nd the other patient was cooled in the 24-hour groupnd also received rt-PA. The third patient with hemor-hagic transformation was cooled as part of the 12-hourroup but did not receive rt-PA. In this case of left mainoronary artery infarction, bleeding involved the left basalanglia with extension into the external capsule but wasonfatal and was not considered a serious adverse event.Two groin site hematomas were reported after femoral

ein puncture; neither was considered serious. rt-PA wasiven to 1 of these patients but not to the other. Otherleeding complications were minor (i.e., not consideredignificant adverse events) and were considered “notelated” or “probably not related” to hypothermia: hema-uria due to Foley catheter–associated trauma (n � 2) andastrointestinal bleeding (n � 2).Average platelet counts decreased slightly in all 3

roups of patients. For patients exposed to �1.5 hours ofypothermia, platelet counts decreased from 222.0k �

0.5k (range, 181k-262k) to 212.0k � 29.1k (range, 182k-40k). For patients in the 12-hour group, platelet countsecreased from 219.1k � 63.9k (range, 149k-372k) to99.5k � 50.5k (range, 124k-199k). For patients in the4-hour group, platelet counts decreased from 198.8k �

5.3k (range, 125k-250k) to 186.4k � 47.8k (range, 108k-25k). However, thrombocytopenia was not consideredo be an adverse event for any patient. The infectionseported through day 30 for the study patients are sum-arized in Table 5, sorted by duration of hypothermia.

Discussion

The ICTuS study was a development/feasibility studyesigned to treat acute stroke patients with endovascularypothermia. Throughout the course of the trial, severalrotocol refinements allowed us to derive the final pro-

ocol. Cooling performance was enhanced by prophylac-ic thermoregulatory suppression; when a conservativentishivering regimen (i.e., surface warming blanket and

ed through day 30

Duration of hypothermia

12 hrs cooling, 12hrs rewarming

(n � 9)

24 hrs cooling, 12hrs rewarming

(n � 6)

No. % No. %

2 22% 0 0%0 0% 1 16%3 33% 1 16%1 11% 0 0%2 22% 0 0%

report

3)

%

33%0%

33%0%0%

0 0% 1 16%

sw1riia(tcsr

eocfiamstcpct(isn

cwtmdpofdagem0roefab

esia

iapa

SBNUS

SPDMS

SHSOP

SFRUH

SKMOO

CPKUS

SJJID

SJRIS

INTRAVASCULAR COOLING IN THE TREATMENT OF STROKE 113

mall boluses of meperidine in response to shivering)as used, the lowest achievable temperature was 35.6 �

.0°C. In contrast, when a more proactive antishiveringegimen was used, in which oral buspirone plus a load-ng dose of IV meperidine followed by a continuousnfusion was administered prophylactically, the lowestchievable temperature in this study was 33.7 � 0.7°Cunpaired t-test, P � .01). This final protocol must now beested in a larger phase 1/2 trial including appropriateontrols. A very similar protocol was recently used in atudy of acute stroke patients, with acceptable safetyesults.9

The incidence and nature of the reported adversevents were consistent with previously published studiesf hypothermia in similar patient populations.1-6,8 In-reasing the duration of hypothermia administrationrom 12 hours to 24 hours did not appear to increase thencidence or severity of adverse effects. The 11 significantdverse events reported in 9 patients reflect the seriousedical conditions of the enrolled patients, and only 2

ignificant adverse events were attributed to the hypo-hermia protocol. However, without a control group, noonclusions can be drawn about the safety profile of thisrotocol, and a further safety trial is necessary, includingontrols, to ensure safety. The 3 hemorrhagic transforma-ions are particularly worrisome, although the incidence3/18; 17%) is consistent with the natural history ofschemic stroke, especially if rt-PA used. A larger safetytudy of endovascular cooling combined with rt-PA iseeded to ensure the safety of this combination.Numerous drugs alter thermoregulatory control, in-

luding most anesthetics and narcotics. Meperidine,hich decreases the shivering threshold twice as fast as

he vasoconstriction threshold through an unknownechanism, is the most important and clinically relevant

rug for inhibiting thermoregulatory responses.17 A me-eridine dose of 25 mg IV decreased the shivering thresh-ld by 2°C, with no effect on alertness or respiratoryunction. To prevent substantial sedation and respiratoryepression associated with high-dose meperidine, othergents may be combined with meperidine to produce areater antishivering effect. Buspirone reduced the shiv-ring threshold to 35.0°C � 0.8°C, whereas high-doseeperidine reduced the shivering threshold to 33.4°C �

.3°C.13 The combination of small doses of these 2 drugseduced the shivering threshold to 33.4°C � 0.7°C withnly minimal sedation. The mechanism of the synergisticffects of buspirone and meperidine is unknown. Weound that the combination of buspirone, meperidine,nd modest surface warming with a forced-air warminglanket produced the optimal shivering suppression.The ICTuS study has enabled the derivation of an

ndovascular cooling protocol that may be safe in elderlytroke victims. Using the final version of the protocol,ncluding a proactive antishivering regimen, we were

ble to induce, maintain, and reverse mild hypothermia

n conscious stroke victims without the need for paralysisnd mechanical ventilation to control shivering. A largerhase 1/2 protocol is needed to confirm the safety of thispproach.

Appendix 1: ICTuS Study Sites

ite 01rett Meyer, MDancy Kelly, RNCSD Medical Center

an Diego, CA

ite 02aul Akins, MDeidre Wentworth, RNercy General Hospital

acramento, CA

ite 03elmi Lutsep, MDusie Fisher, RNregon Health Sciences Universityortland, OR

ite 04ahmi Al-Sanani, MDobin Saiki, RNniversity of Texas, Houston Medical Schoolouston, TX

ite 05en Ng, MDelissa Holycross, BScala Neurodiagnostic Centercala, FL

oordinating Centeratrick Lyden, MDaren Rapp, RNCSD Clinical Trial Coordinating Center

an Diego, CA

tudy Safety and Monitoringustin Zivin, MD, PhDulie Jurf, RNND Inc.allas, TX

tudy Sponsorohn Dobak, MDobin Allgren, MD, PhD

NNERCOOL Therapies, Inc.an Diego, CA

References

1. Bernard SA, Buist MD, Jones BM, et al. Treatment ofcomatose survivors of out-of-hospital cardiac arrest with

induced hypothermia. New Engl J Med 2002;346:557-563.

1

1

1

1

1

1

1

1

P.D. LYDEN ET AL.114

2. The Hypothermia After Cardiac Arrest Study Group.Mild therapeutic hypothermia to improve the neurologicoutcome after cardiac arrest. New Engl J Med 2002;346:549-556.

3. Kammersgaard L, Rasmussen BH, Jorgensen HS, et al.Feasibility and safety of inducing modest hypothermiain awake patients with acute stroke through surfacecooling: A case-control study. Stroke 2000;31:2251-2256.

4. Krieger D, De Georgia M, Abou-Chebl A, et al. Coolingfor acute ischemic brain damage (COOL AID). Stroke2001;32:1847-1854.

5. Schwab S, Schwarz S, Spranger M, et al. Moderate hy-pothermia in the treatment of patients with severe mid-dle cerebral artery infarction. Stroke 1998;29:2461-2466.

6. Steiner T, Freide T, Aschoff A, et al. Effect and feasibilityof controlled rewarming after moderate hypothermia instroke patients with malignant infarction of the middlecerebral artery. Stroke 2001;32:2833-2835.

7. Zwefler RM, Voorhees ME, Mahmood MA, et al. Induc-tion and maintenance of mild hypothermia by surfacecooling in nonintubated subjects. J Stroke CerebrovascDis 2003;12:237-243.

8. Schwab S, Georgiadis D, Berrouschot J, et al. Feasibilityand safety of moderate hypothermia after massive hemi-spheric infarction. Stroke 2001;32:2033-2035.

9. De Georgia MA, Krieger DW, Abou-Chebl A, et al.

Cooling for acute ischemic brain damage (COOL AID): A

feasibility trial of endovascular cooling. Neurology 2004;63:312-317.

0. Lyden P, Brott T, Tilley B, et al and theNINDS TPAStroke Study Group. Improved reliability of the NIHstroke scale using video training. Stroke 1994;25:2220-2226.

1. Mahoney FT, Barthel DW. Functional evaluation: Barthelindex. Md State Med J 1965;14:61-65.

2. Rankin J. Cerebral vascular accidents in patients over theage of 60: Prognosis. Scott Med J 1957;2:200-215.

3. Mokhtarani M, Mahgoub AN, Morioka M, et al. Buspi-rone and mederidine synergistically reduce the shiver-ing threshold. Anesth Anal 2001;93:1233-1239.

4. NINDS rt-PA Stroke Study Group. Tissue plasminogenactivator for acute ischemic stroke. N Engl J Med 1995;333:1581-1587.

5. Mian NZ, Bayly R, Schreck DM, et al. Incidence of deepvenous thrombosis associated with femoral venous cath-eterization. Acad Emerg Med 1997;4:1118-1121.

6. Johnston KC, Li JY, Lyden PD, et al. Medical and neu-rological complications of ischemic stroke: Experiencefrom the RANTTAS trial. RANTTAS Investigators.Stroke 1998;29:447-453.

7. Alfonsi P, Sessler DI, Du Manoir B. The effects of me-peridine and sufentanil on the shivering threshold in

postoperative patients. Anesthesiology 1998;89:43-48.