Survival without brain damage after clinical death of 60-120 mins in dogs using suspended animation by profound hypothermia.

Abstract

This study explored the limits of good outcome of brain and organism achievable after cardiac arrest (no blood flow) of 60-120 mins, with preservation (suspended animation) induced immediately after the start of exsanguination cardiac arrest.
Prospective experimental comparison of three arrest times, without randomization.
University research laboratory.
Twenty-seven custom-bred hunting dogs (17-25 kg).
Dogs were exsanguinated over 5 mins to cardiac arrest no-flow of 60 mins, 90 mins, or 120 mins. At 2 mins of cardiac arrest, the dogs received, via a balloon-tipped catheter, an aortic flush of isotonic saline at 2 degrees C (at a rate of 1 L/min), until tympanic temperature reached 20 degrees C (for 60 mins of cardiac arrest), 15 degrees C (for 60 mins of cardiac arrest), or 10 degrees C (for 60, 90, or 120 mins of cardiac arrest). Resuscitation was by closed-chest cardiopulmonary bypass, postcardiac arrest mild hypothermia (tympanic temperature 34 degrees C) to 12 hrs, controlled ventilation to 20 hrs, and intensive care to 72 hrs.
We assessed overall performance categories (OPC 1, normal; 2, moderate disability; 3, severe disability; 4, coma; 5, death), neurologic deficit scores (NDS 0-10%, normal; 100%, brain death), regional and total brain histologic damage scores at 72 hrs (total HDS >0-40, mild; 40-100, moderate; >100, severe damage), and morphologic damage of extracerebral organs. For 60 mins of cardiac arrest (n = 14), tympanic temperature 20 degrees C (n = 6) was achieved after flush of 3 mins and resulted in two dogs with OPC 1 and four dogs with OPC 2: median NDS, 13% (range 0-27%); and median total HDS, 28 (range, 4-36). Tympanic temperature of 15 degrees C (n = 5) was achieved after flush of 7 mins and resulted in all five dogs with OPC 1, NDS 0% (0-3%), and HDS 8 (0-48). Tympanic temperature 10 degrees C (n = 3) was achieved after flush of 11 mins and resulted in all three dogs with OPC 1, NDS 0%, and HDS 16 (2-18). For 90 mins of cardiac arrest (n = 6), tympanic temperature 10 degrees C was achieved after flush of 15 mins and resulted in all six dogs with OPC 1, NDS 0%, and HDS 8 (0-37). For 120 mins of cardiac arrest (n = 7), three dogs had to be excluded. In the four dogs within protocol, tympanic temperature 10 degrees C was achieved after flush of 15 mins. This resulted in one dog with OPC 1, NDS 0%, and total HDS 14; one with OPC 1, NDS 6%, and total HDS 20; one with OPC 2, NDS 13%, and total HDS 10; and one with OPC 3, NDS 39%, and total HDS 22.
In a systematic series of studies in dogs, the rapid induction of profound cerebral hypothermia (tympanic temperature 10 degrees C) by aortic flush of cold saline immediately after the start of exsanguination cardiac arrest-which rarely can be resuscitated effectively with current methods-can achieve survival without functional or histologic brain damage, after cardiac arrest no-flow of 60 or 90 mins and possibly 120 mins. The use of additional preservation strategies should be pursued in the 120-min arrest model.

Full text

Page 1

Neurologic Critical Care
Survival without brain damage after clinical death of 60–120 mins
in dogs using suspended animation by profound hypothermia*
Wilhelm Behringer, MD; Peter Safar, MD; Xianren Wu, MD; Rainer Kentner, MD;
Ann Radovsky PhD; Patrick M. Kochanek, MD; C. Edward Dixon, PhD; Samuel A. Tisherman, MD
I
mon, and exsanguination cardiac arrest
(CA), which is no-flow and rare. CA is the
topic of this study. Civilian trauma pa-
tients and military combat casualties
with penetrating (often repairable)
n considering resuscitation from
severe hemorrhage, one must dif-
ferentiate between hemorrhagic
shock, which is low-flow and com-
trunkal injuries exsanguinate rapidly to
CA. Conventional resuscitation attempts
are futile, and survival rates are near zero
(1–4). For such unresuscitable condi-
tions, since 1984, Safar and Bellamy have
recommended research into “suspended
animation for delayed resuscitation.”
This they have defined as “induction of
preservation of the organism within the
first 5 min of CA (no-flow) for transport
and surgical hemostasis during clinical
death, to be followed by delayed resusci-
tation to survival without brain damage”
(4).
Treatment induced before arrest (pro-
tection) and maintained during arrest
(preservation) is more likely to mitigate
postischemic brain damage than when
induced after arrest (resuscitation) (5, 6).
Suspended animation is preservation-
resuscitation with use of drugs or hypo-
thermia. Using systematic studies of ex-
sanguination CA in a reproducible dog
outcome model with induction of preser-
vation by aortic flush at 2 mins CA, of
saline at 24°C, via a balloon-tipped cath-
eter (7–10), we obtained disappointing
Objectives: This study explored the limits of good outcome of
brain and organism achievable after cardiac arrest (no blood flow)
of 60–120 mins, with preservation (suspended animation) in-
duced immediately after the start of exsanguination cardiac ar-
rest.
Design: Prospective experimental comparison of three arrest
times, without randomization.
Setting: University research laboratory.
Subjects: Twenty-seven custom-bred hunting dogs (17–25 kg).
Interventions: Dogs were exsanguinated over 5 mins to cardiac
arrest no-flow of 60 mins, 90 mins, or 120 mins. At 2 mins of
cardiac arrest, the dogs received, via a balloon-tipped catheter,
an aortic flush of isotonic saline at 2°C (at a rate of 1 L/min), until
tympanic temperature reached 20°C (for 60 mins of cardiac
arrest), 15°C (for 60 mins of cardiac arrest), or 10°C (for 60, 90,
or 120 mins of cardiac arrest). Resuscitation was by closed-chest
cardiopulmonary bypass, postcardiac arrest mild hypothermia
(tympanic temperature 34°C) to 12 hrs, controlled ventilation to 20
hrs, and intensive care to 72 hrs.
Measurements and Main Results: We assessed overall perfor-
mance categories (OPC 1, normal; 2, moderate disability; 3, severe
disability; 4, coma; 5, death), neurologic deficit scores (NDS
0–10%, normal; 100%, brain death), regional and total brain
histologic damage scores at 72 hrs (total HDS >0–40, mild;
40–100, moderate; >100, severe damage), and morphologic dam-
age of extracerebral organs. For 60 mins of cardiac arrest (n ?
14), tympanic temperature 20°C (n ? 6) was achieved after flush
of 3 mins and resulted in two dogs with OPC 1 and four dogs with
OPC 2: median NDS, 13% (range 0–27%); and median total HDS,
28 (range, 4–36). Tympanic temperature of 15°C (n ? 5) was
achieved after flush of 7 mins and resulted in all five dogs with
OPC 1, NDS 0% (0–3%), and HDS 8 (0–48). Tympanic temperature
10°C (n ? 3) was achieved after flush of 11 mins and resulted in
all three dogs with OPC 1, NDS 0%, and HDS 16 (2–18). For 90
mins of cardiac arrest (n ? 6), tympanic temperature 10°C was
achieved after flush of 15 mins and resulted in all six dogs with
OPC 1, NDS 0%, and HDS 8 (0–37). For 120 mins of cardiac arrest
(n ? 7), three dogs had to be excluded. In the four dogs within
protocol, tympanic temperature 10°C was achieved after flush of
15 mins. This resulted in one dog with OPC 1, NDS 0%, and total
HDS 14; one with OPC 1, NDS 6%, and total HDS 20; one with OPC
2, NDS 13%, and total HDS 10; and one with OPC 3, NDS 39%, and
total HDS 22.
Conclusions: In a systematic series of studies in dogs, the
rapid induction of profound cerebral hypothermia (tympanic tem-
perature 10°C) by aortic flush of cold saline immediately after the
start of exsanguination cardiac arrest—which rarely can be re-
suscitated effectively with current methods—can achieve sur-
vival without functional or histologic brain damage, after cardiac
arrest no-flow of 60 or 90 mins and possibly 120 mins. The use of
additional preservation strategies should be pursued in the 120-
min arrest model. (Crit Care Med 2003; 31:1523–1531)
KEY WORDS: cardiac arrest; hypothermia; hemorrhage; resusci-
tation; cerebral ischemia; cardiopulmonary bypass
*See also p. 1592.
From the Safar Center for Resuscitation Re-
search (WB, PS, XW, RK, AR, PMK, CED, SAT),
Department of Anesthesiology/Critical Care Medi-
cine (PS), Department of Surgery (SAT), Depart-
ment of Pediatrics (PMK), and Department of Neu-
rosurgery (CED), University of Pittsburgh, Pittsburgh,
PA.
Supported, in part, by the U.S. Department of
Defense, Office of Naval Research, grant N00014-97-
1-1064; by the U.S. Army’s MRMC/TATRC, grant
N00014-99-1-0765; and by the Cardeon Corporation,
which provided the flush catheter.
Copyright © 2003 by Lippincott Williams & Wilkins
DOI: 10.1097/01.CCM.0000063450.73967.40
1523Crit Care Med 2003 Vol. 31, No. 5

Page 2

outcome results with a series of mecha-
nism-specific pharmacologic therapies
(11–15). The exception was the antioxi-
dant tempol, which improved functional
outcome (15). In contrast, lowering the
temperature of the flushed saline to 2°C
and progressively increasing the flush
volume, starting the flush at 2 mins of
normothermic exsanguination CA, we
could decrease brain (tympanic mem-
brane) temperature (Tty) to around 34°C,
which preserved brain viability during CA
of 15 mins (7) and 20 mins (8), and to
around 28°C, which preserved brain via-
bility for 30 mins (9). The present study is
an extension of these systematic efforts to
maximize the duration of CA (no-flow)
from which resuscitation to survival can
be achieved without vital organ system
damage (10). Attempting to extend this
maximal CA period from 30 to 120 mins
is called for by the fact that transport and
surgical hemostasis in patients with trau-
matic exsanguination to CA would re-
quire such prolonged preservation, par-
ticularly in military combat scenarios.
Protective-preservative hypothermia,
induced and reversed with cardiopulmo-
nary bypass (CPB), is clinically used for
some elective operations on heart or
brain but has not been evaluated yet for
emergency scenarios as in this study.
Elective therapeutic hypothermia has
been shown to protect the brain and
whole organism in animals or patients for
up to 15 mins of CA at brain temperature
of about 35°C (mild hypothermia) (16,
17), for up to 20 mins of CA at about 30°C
(moderate hypothermia) (18), for up to
30 mins of CA at about 20°C (deep hypo-
thermia) (19), for up to 60–150 mins of
CA at 5–10°C (profound hypothermia)
(20–27), and perhaps even for longer CA
with ultraprofound hypothermia (28–
30). The normal brain is not damaged by
temperatures lowered to 5–10°C (31) but
can be damaged by temperatures below
5°C (32, 33). In most of the previously
mentioned studies of protective-preserva-
tive hypothermia for elective prolonged
CA (23–30), induction of hypothermia
was with CPB, before induction of CA and
without total exsanguination; also, eval-
uation of cerebral function and histology
was not quantitative as in our present
study.
The experiments reported in this arti-
cle are the first systematic explorations of
emergency measures aimed at maximiz-
ing the reversible CA no-flow duration.
The method should ultimately be induc-
ible for patients also outside hospitals.
Ours is the only group experimenting
with this suspended animation approach,
which includes early and rapid induction
of preservation with aortic flush and in-
tensive care life support for 72 hrs to give
the ischemic anoxic encephalopathy time
to mature. The objective of this study was
to simulate the scenario of rapid exsan-
guination to death from a laceration in
the aorta or vena cava and to determine,
for the first time, the longest no-flow
period from which resuscitation to com-
plete recovery can be accomplished with
the aid of CPB. This study is also the first
to use a single aortic saline flush imme-
diately after the start of CA (no-flow), to
include preservation and long-term in-
tensive care to outcome evaluation in
terms of function and semiquantitative
histologic brain damage. We hypothe-
sized (10) that preservation during CA 60,
90, or 120 mins (no-flow) requires Tty
10°C to achieve intact survival without
histologic brain damage.
MATERIALS AND METHODS
This study was approved by the Institu-
tional Animal Care and Use Committee of the
University of Pittsburgh and the Department
of Defense and followed national guidelines for
the treatment of animals. All experiments
were conducted by the same team between
May 2000 and January 2001, in mixed se-
quence, without randomization. The protocol
called for exsanguination to CA: in study A
with 60 mins of no-flow, comparing three pre-
servative levels of hypothermia (Tty 20, 15, or
10°C), and in study B with 90 mins vs. 120
mins of no-flow at Tty 10°C. Resuscitation was
with CPB, mild hypothermia to 12 hrs, con-
trolled ventilation to 20 hrs, and intensive care
to final outcome evaluation at 72 hrs. At CA 2
mins, the dogs received a flush into the aorta
with saline at 2°C, in study A until Tty reached
20°C (n ? 6), 15°C (n ? 5), or 10°C (n ? 3)
for CA 60 mins, and in study B until Tty
reached 10°C for CA 90 mins (n ? 6) or 120
mins (n ? 7, of which 3 had to be excluded, as
discussed in the Results).
Preparation. We studied 27 custom-bred
hunting dogs (17–25 kg body weight, age
8–12 months, simulating young healthy
trauma victims). Details of preparation and
model have been described (7–9, 13–15).
Briefly, after premedication with ketamine,
orotracheal intubation, and anesthesia with
halothane and N2O/oxygen (1:1), temperature
probes were inserted for measuring Tty and
esophageal (Tes) and rectal temperatures (Tr).
In pilot experiments during flush, brain tissue
temperature decreased more rapidly than Tty,
but the two equilibrated rapidly within ?1°C.
Intravenous maintenance fluid was with dex-
trose 5% in 0.45% NaCl. The left femoral
artery was cannulated for monitoring of arte-
rial pressure. A pulmonary artery catheter was
inserted to monitor pressure, cardiac output,
and temperature (Tpa). A prototype balloon
catheter (8 Fr), with one hole at the tip of the
catheter, was advanced via the right femoral
artery into the aorta for arterial bleeding and
for the aortic flush. The right external jugular
vein was cannulated with a multiple-holed
cannula (18 Fr), which was advanced to the
level of the right atrium for venous bleeding
and for venous return to the CPB system.
Arterial and central venous pressures and the
electrocardiogram were continuously re-
corded. Arterial and mixed venous blood gases,
hemoglobin, hematocrit, sodium, potassium,
glucose, and lactate were measured at regular
intervals. Blood gases were controlled as mea-
sured at normothermia (alpha-stat strategies).
Samples for liver function (glutamyl oxaloace-
tic transaminase, ?-glutamyl transpeptidase,
and bilirubin in serum) and kidney function
(creatinine in serum and urine) were taken at
baseline, 24 hrs, and 72 hrs. Just before start
of the insult, Tty was controlled at 37.5 ?
0.1°C by heating blanket and lamp.
Insult. After two baseline measurements,
heating devices, intravenous fluids, and halo-
thane were discontinued, while the dogs were
weaned to spontaneous breathing of N2O/
oxygen (2:1) via a T-tube. When the canthal
reflex returned (as an indication of very light
anesthesia), hemorrhage was initiated. Over a
5-min period, the dogs were bled via the arte-
rial and venous cannulae (simulating trau-
matic laceration), and the blood was collected
in bags with sodium citrate anticoagulant for
later reinfusion. Hemorrhage was controlled
to mean arterial pressure (MAP) 20 mm Hg at
4 mins. At 5 mins, to ensure zero blood flow,
ventricular fibrillation was induced with one
or more subcutaneous transthoracic shocks of
110 V AC, to fully control the onset of circu-
latory arrest. Total arrest (no-flow) time was
60 mins in study A and 90 or 120 mins in
study B.
Aortic Flush. Two minutes after the onset
of CA, the balloon of the aortic catheter,
placed in the abdominal aorta, was inflated
with 1.5 mL of saline, known to occlude the
aorta (9). Saline at 2°C then was flushed into
the aorta at a rate of 1 L/min by using a roller
pump (care was taken to avoid air entering the
system). In study A, the flush was stopped
when Tty reached 20°C (n ? 6), 15°C (n ? 5),
or 10°C (n ? 3). In study B, for CA 90 mins,
the flush was stopped when Tty reached 10°C
(n ? 6). For CA 120 mins, a pilot experiment
had shown the hind legs with rigor mortis at
start of resuscitation, and the rectum was ne-
crotic at necroscopy. Therefore, in study B, for
CA 120 mins, the balloon was first placed in
the thoracic aorta until Tty reached 10°C and
then deflated and pulled back into the femoral
artery, continuing the flush until Tr reached
20°C (n ? 7). After the flush, during CA, the
1524Crit Care Med 2003 Vol. 31, No. 5

Page 3

aortic catheter was replaced with a short arte-
rial CPB cannula (7 or 8 Fr).
During CA, the head and neck were im-
mersed in ice water to prevent the spontane-
ous rewarming by 3–5°C we have seen previ-
ously during CA of 60–120 mins without ice
water immersion. In two pilot experiments,
laryngeal-pharyngeal edema occurred after
the whole neck was immersed in ice water ?1
hr; therefore, in study B only the head was
immersed in ice water and neck edema did not
occur.
Resuscitation. After CA no-flow of 60, 90,
or 120 mins, reperfusion was with CPB, be-
cause standard cardiopulmonary resuscitation
cannot reliably achieve restoration of sponta-
neous circulation (ROSC) from CA ?12 mins
of no-flow (6, 16, 17). The CPB circuit was
primed with 400 mL of Dextran 40 10% plus
Ringer’s solution (1:1). Sodium bicarbonate (2
mEq/kg) and heparin (1500 units) were added.
Just before the start of CPB, additional sodium
bicarbonate (1 mEq/kg) was injected into the
circuit. The dogs were paralyzed with pancu-
ronium (0.1 mg/kg intravenously). The tem-
perature of the water bath of the CPB heat
exchanger was set to 5°C above Tty, until Tty
reached 34°C. CPB was started with a flow of
50 mL·kg?1·min?1for Tty ?20°C, increased
to 75 mL·kg?1·min?1for Tty 21°C–30°C, and
increased to 100 mL·kg?1·min?1for Tty
?30°C. Reinfusion of all shed blood was ti-
trated to achieve a central venous pressure of
10–15 mm Hg. Repetitive doses of epineph-
rine (0.01 mg/kg) were given intra-arterially
as necessary to increase MAP to 60 mm Hg
during Tty ?20°C, to 80 mm Hg during Tty
21–30°C, and to 100 mm Hg during Tty
?30°C. When Tpa reached 32°C, defibrillation
attempts were with external DC counter-
shocks of 150 J, increased by 50 J for repeated
shocks. Oxygen flow through the oxygenator
was adjusted to keep PaCO2at 30–35 mm Hg
and PaO2?100 mm Hg. During CPB of 2 hrs,
controlled ventilation was with 100% oxygen
at a rate of eight to ten inflations per minute.
The intravenous fluids were restarted with a
flow of 100 mL/hr. A base deficit of ?6.0
mEq/L was corrected with sodium bicarbon-
ate. When ROSC was established, a norepi-
nephrine infusion was titrated intravenously
to achieve a brief hypertension of MAP ?150
mm Hg (9). We include in standard protocols
hypertensive reperfusion, which improves ce-
rebral blood flow and outcome (34). Thereaf-
ter, MAP was controlled at 90–150 mm Hg.
The CPB flow rate for assisted circulation was
reduced to 75 and 50 mL·kg?1·min?1and
stopped at 120 mins. During CPB, activated
clotting times were maintained at ?300 secs
with additional heparin as needed.
Intensive Care. After weaning from CPB
assist at 2 hrs, controlled ventilation was con-
tinued to 20 hrs with N2O/oxygen 1:1 for an-
algesia. Paralysis was maintained with inter-
mittent doses of pancuronium. To prevent
stress, fentanyl boluses (5–10 ?g/kg) were
given intravenously whenever signs of possible
distress (mydriasis, tachycardia, or hyperten-
sion) occurred. In pilot experiments without
paralysis, there were no escape movements
with this regimen (9). Hypotension (MAP ?90
mm Hg) was treated with titrated intravenous
infusion of Ringer’s solution or norepineph-
rine. Severe hypertension (MAP ?150 mm
Hg) was controlled with titrated intravenous
boluses of labetalol or hydralazine. Standard
intensive care included airway suctioning, pe-
riodic deep lung inflations, and position
change (rotation). The dogs received cefazolin
every 8 hrs for infection prophylaxis. At 20–24
hrs, paralysis was reversed to spontaneous
breathing with neostigmine plus atropine. The
dogs were extubated when they were able to
maintain normal blood gas values during
spontaneous breathing and after upper airway
reflexes had returned. The catheters were then
removed under brief, light N2O-halothane an-
esthesia by mask. When awake and stable, the
dog was transferred to a stepdown intensive
care unit to 72 hrs, with oxygen by mask,
continuous monitoring of pulse rate, and ar-
terial oxygen saturation by technicians and
critical care physicians. Seizures, running
movements, opisthotonos, or spontaneous
tachypnea were controlled with titrated doses
of diazepam (0.2–0.3 mg/kg intravenously) as
needed. Tty was controlled at 34°C with exter-
nal cooling and warming for the first 12 hrs
after start of CPB and at 37.5°C until 72 hrs.
The maintenance intravenous fluid was dex-
trose 5% in NaCl 0.45% until 24 hrs and
dextrose 10% in NaCl 0.45% thereafter for
supply of energy—until the dog was able to
eat and drink. Optimal blood glucose concen-
trations after prolonged CA are unknown (35,
36).
Outcome Evaluation. Function and cere-
bral morphologic changes were evaluated as
described before (6, 34, 37–40). Briefly, per-
formance was evaluated according to overall
performance categories (OPC 1, normal; 2,
moderate disability; 3, severe disability; 4,
coma; and 5, death or brain death). Neurologic
function was evaluated as neurologic deficit
scores (NDS 0–10%, normal; 100% ? brain
death). OPC and NDS were evaluated every 8
hrs after extubation. Final evaluations (72 hrs)
were independently determined and agreed
upon by two team members. Attempts were
made to discontinue any sedation ?4 hrs be-
fore final evaluations. If necessary, sedation
was reversed with flumazenil.
After final outcome evaluation at 72 hrs,
for morphologic studies (37), the dogs were
reanesthetized as before, the left hemithorax
was opened, and the proximal descending
aorta was ligated. A large-bore cannula was
inserted proximal to the ligature. The dogs
then were killed by infusing into the aortic
arch approximately 2 L of paraformaldehyde
(4%, pH 7.4). A complete necropsy was per-
formed, and samples of extracerebral organs
were taken for histologic examination. Macro-
scopic scoring was performed of damage in
gut and heart (mild vs. moderate vs. severe
hemorrhage, any necrosis).
One hour after perfusion fixation, the brain
was removed. After 3-mm thick slices were
cut, the same six slices of each brain were
paraffin embedded, cut into sections 4 mi-
crons thick, and stained with hematoxylin-
eosin-phloxine (37). Using light microscopy,
the same pathologist (AR), unaware of treat-
ment, group assignments, and hypotheses,
scored 19 distinct anatomical brain regions for
severity and extent of ischemic neuronal
changes (shrunken eosinophilic neurons with
pyknotic nuclei), infarcts, and edema, as de-
scribed previously (37, 38). The total brain
histologic damage score (HDS) was the sum of
all area scores. A total HDS between zero and
about 40 has usually correlated with normal
or minimally impaired function, about 40–
100 represented moderate damage, and total
HDS ?100 indicated severe damage (37, 38).
Extracerebral variables were monitored as de-
scribed in the Results.
For exploration of cognitive function re-
covery, three dogs with OPC ? 1 and NDS
0–10% (normal) at 72 hrs—one of study A
after CA 60 mins at Tty 20°C, one of study B
after CA 90 mins at Tty 10°C, and one normal
dog without CA—were evaluated over 6
months for their ability to learn a spatial ver-
sion of a successive reversal learning task,
modified from Head et al. (41). The dogs were
initially habituated to a test chamber. On one
wall of the chamber were two head holes that
could be covered or opened by a sliding door.
Outside of each head hole was a block that
covered the reward. The dogs were trained to
find a food reward based on the spatial posi-
tion (left vs. right) that was the reverse of the
preceding trial. The reward was alternatively
hidden under the left or right block and a
correct response was counted if the dog chose
the side with the reward. We measured the
number of sessions, of ten consecutive trials
each, needed to twice achieve eight of ten
correct responses.
Statistical Analysis. Dogs that either did
not meet protocol criteria or died from extra-
cerebral causes before 72 hrs were excluded
from outcome analysis. Brain death as an out-
come was included if the study process satis-
fied protocol. Data are given as mean and SD, if
normally distributed, and otherwise as median
and range. This study was an exploratory one
not depending on statistical group differences,
and in study A, we did not perform any statis-
tical group comparisons, since one group con-
sisted of only three dogs. Merely for the pur-
pose of complete information, we used in
study B the independent samples t-test or the
Mann-Whitney U test to compare continuous
variables (physiologic variables, NDS, HDS)
and the chi-square test for trend to test for
differences in proportions of OPC values be-
tween groups. All data were computed with
SPSS for Windows (release 8.0; Chicago, IL) or
NCSS for Windows (UT). We considered p ?
.05 to be statistically significant.
1525Crit Care Med 2003 Vol. 31, No. 5

Page 4

RESULTS
For both studies, a total of 27 dogs
were exsanguinated to CA. In study A
with CA 60 mins (n ? 14) and in study B
with CA 90 mins (n ? 6), all dogs sur-
vived to 72 hrs in protocol. In study B
with CA 120 mins (n ? 7), three dogs
failed to survive to 72 hrs and had to be
excluded because one developed pulmo-
nary edema during CPB, hemorrhagic di-
arrhea, anemia, and hemorrhagic lungs
and was killed at 23 hrs; one had the flush
delayed due to technical pump error; and
one died at 25 hrs unrecognized in the
stepdown unit after premature extuba-
tion due to human error. Thus, only four
of the seven dogs in the CA 120-min
group survived to final evaluation at 72
hrs. One dog of study A after CA 60 mins
at Tty 20°C, and one dog of study B after
CA 90 mins at Tty 10°C, although within
protocol, provided only functional out-
come data because they were kept alive
for cognitive function testing at 6
months, leaving 22 of the 27 dogs for
morphologic evaluation at 72 hrs.
Heart rate, MAP, and arterial PO2,
PCO2, hematocrit, base excess, sodium,
potassium, glucose, and lactate—at base-
line and at resuscitation time 6 hrs—
were within normal ranges, except in
study B at 6 hrs, where the blood glucose
values in the CA 120-min group were
statistically higher (median, 316 mg/dL;
range, 289–363) than in the CA 90-min
group (median, 227 mg/dL; range, 107–
253; p ? .01).
Flush and Temperatures. Flush vol-
umes and durations needed to achieve
the target Tty are given in Table 1. After
we stopped the aortic flush, Tty decreased
spontaneously slightly further, in study A
with CA 60 mins to a lowest Tty of 18.3 ?
1.8°C (range, 15.1–20.0) in the 20°C
group; to 14.4 ? 0.3°C (14.2–14.9) in the
15°C group; and to 10°C, 9.7°C, and
9.8°C in the 10°C group (Fig. 1A). In
study B with CA 90 mins, the lowest Tty
was 9.2 ? 1.0°C (7.5–10.0; Fig. 1B), and
with CA 120 mins the lowest Tty was 6.9
? 0.8°C (5.9–7.9; Fig. 1B). Lowest Tes
(core T) ranged between 15.1 and 24.8°C
in the Tty 20°C group, between 16.3°C
and 22.8°C in the Tty 15°C group, and
between 6.1 and 15.3°C in the Tty 10°C
groups. Lowest Tr was approximately 35–
37°C, except for the CA 120-min dogs
(with cold flush into the femoral artery),
in which lowest Tr ranged between 20.0
and 24.6°C. In one dog, the flush was
stopped when 15 L flush volume was con-
sumed, with Tr 24.6°C.
Resuscitation. During reperfusion
with CPB, the time required to increase
core temperature (Tpa) to 32°C depended
on the depth of hypothermia at the end of
CA (Table 1). In both studies, once Tpa
reached 32°C, ROSC was achieved after
one to three countershocks (Table 1). In
study A, one dog in the 20°C group spon-
taneously developed QRS complexes at
the start of CPB and ROSC at 5 mins after
the start of CPB. The amount of epineph-
rine required to keep MAP in protocol
during CPB did not differ between
groups, nor did the amounts of norepi-
nephrine and bicarbonate required after
ROSC (Table 1). There were no signifi-
cant group differences in the highest
MAP and the duration of the brief in-
duced hypertension, which ranged be-
tween 1 and 14 mins (Table 1).
Complications. In addition to the
complications that caused three exclu-
sions (described previously), there were
complications in four of the included
dogs: One of the five included dogs with
CA 60 mins at 15°C developed increased
pulmonary artery occlusion pressure and
pulmonary edema while on controlled
ventilation. The pulmonary edema was
managed with increasing positive end-
expiratory pressure; arterial PO2re-
mained ?200 mm Hg. One dog after CA
60 mins developed fever in the stepdown
unit at 32 hrs, which was reversed with
acetaminophen by mouth. One dog of
this group had hemorrhagic diarrhea
during the observation phase. One dog,
after CA 90 mins at 10°C, had hemor-
rhagic diarrhea during the observation
phase; one dog of the same group devel-
oped tachycardic atrial fibrillation 1 hr
after the start of CPB, which was resistant
to amiodarone and diltiazem but could be
terminated with one synchronized coun-
tershock (50 J) at 24 hrs.
Overall and Cerebral Outcome. OPC,
NDS, and total brain HDS at 72 hrs are
shown in Figure 2; regional HDS is
shown in Figure 3.
In study A, after CA 60 mins at Tty
20°C, two dogs achieved OPC 1 with NDS
0% and 1% (normal). In one of these
dogs, total brain HDS was 28, and the
other dog was kept for cognitive function
testing. Four dogs achieved OPC 2 due to
various degrees of motor impairment of
the hind legs, including inability to stand
and walk, which resulted in NDS 6–27%;
Table 1. Aortic flush volume and duration needed to achieve target tympanic temperature: Requirements for restoration of spontaneous circulation (ROSC)
and hypertensive bout
Study A (60 Mins CA) Study B (Tty 10°C)
Tty 20°C (n ? 6) Tty 15°C (n ? 5) Tty 10°C (n ? 3)90 Mins (n ? 6) 120 mins (n ? 4)
Flush volume, mL/kg
Flush duration, min:sec
Time to reach Tpa 32°C, min
Countershocks, total number
Countershocks, total energy, J
Total bicarbonate, mEq
Total epinephrine, mg
Total norepinephrine, mg
Hypertensive bout: peak MAP, mm Hg
Hypertensive bout start, mina
Hypertensive bout duration, minb
159
3:30
22
(144–228)
(3:15–5:30)
(11–23)
(0–2)
(0–300)
(165–235)
0.9 (0.2–2.4)
0.96 (0.80–1.76)
175(160–200)
22(9–29)
5 (3–7)
306
7:10
20
(258–373)
(5:19–8:40)
(13–55)
(1–3)
(150–500)
(155–270)
1.0 (0.2–1.8)
1.12 (0.96–2.24)
165(150–200)
30 (19–63)
4(3–5)
430, 469, 546
9:11, 10:55, 12:15 14:33
31, 32, 33
1, 1, 1
150, 150, 150
200, 200, 230
0.6, 1.0, 1.5
1.76, 1.92, 2.88
150, 155, 160
40, 41, 43
2, 4, 6
578 (535–736)
(11:50–15:58) 15:17
(24–56)
(1–1)
(150–150)
(190–280)
1.6 (0.4–3.4)
3.12 (0.8–7.68)
165 (150–175)
54(28–64)
4 (1–14)
666(598–755)
(13:45–16:39)
(29–62)
(1–2)
(150–300)
(150–290)
1.3 (1.1–1.8)
1.20 (1.12–1.28)
165(150–170)
49 (36–67)
4(3–4)
44 43
1111
150
205
150
185
150
220
150
250
CA, cardiac arrest; Tty, tympanic temperature; Tpa, pulmonary artery temperature; MAP, mean arterial pressure.
aStart of hypertensive bout ? time after start of cardiopulmonary bypass;bduration of hypertensive bout ? time with MAP ?150 mm Hg. Data are given
as median (range) or as single values.
1526Crit Care Med 2003 Vol. 31, No. 5

Page 5

they showed normal cerebral perfor-
mance. Their total brain HDS was 4–36.
Histologic evaluation of the spinal cord
did not show any pathology. In the Tty
15°C and 10°C groups, all dogs achieved
OPC 1 and NDS 0% (except that one in
the 15°C group had NDS 3% due to weak
hind legs). Total brain HDS was ?40 in
all dogs in the 15°C and 10°C groups,
except one in the 15°C group with worse
HDS. Histologically normal brains were
seen in one dog in the Tty 20°C group
(HDS 4), two dogs in the 15°C group
(HDS 0 and 6), and one dog in the 10°C
group (HDS 2).
In study B, after CA 90 mins at Tty
10°C (n ? 6), all dogs were functionally
normal (OPC 1 and NDS 0%), with total
brain HDS ? 8 (range, 0–37; n ? 5). One
dog had HDS zero. After CA 120 mins (n
? 4), functional outcomes varied. One
dog achieved OPC 1 with NDS 0% and
total HDS 14. One dog achieved OPC 1
with NDS 6% due to mild disability in the
hind legs and HDS 20. One dog achieved
OPC 2 with NDS 13 due to weakened hind
legs and HDS 10. One dog achieved OPC
3 (poor outcome) with NDS 39 and HDS
22. Total brain HDS were ?40 in all dogs
after CA 90 or 120 mins and also in the
one dog with OPC 3.
Regional brain HDS showed the same
distribution in all groups. Therefore, the
results are summarized for all 22 dogs in
which HDS was evaluated (Fig. 3). In 17
of the 19 regions, the median HDS was
zero. Putamen and caudate nucleus
seemed to be the most vulnerable regions
in this model, as was the case in this
model also with moderate hypothermia.
All scores were the result of scattered
ischemic neurons, except for one dog
with CA 90 mins at 10°C that showed also
edema in the thalamus and dentate nu-
cleus, and another dog of the same group
that had also an infarcted area in the
dentate nucleus.
Cognitive function testing in the one
dog with OPC 1 after CA 60 mins at Tty
20°C, and in the one dog with OPC 1 after
CA 90 mins at 10°C, demonstrated that
both were able to meet the criteria for
successfully learning the cognitive task at
3–6 months after CA. Neither of these
two CA dogs performed worse than the
normal dog without CA.
Extracerebral Outcome. Extracerebral
malfunction was transient, and morpho-
logic changes at 72 hrs were not severe in
both studies (Table 2). Variables reflect-
ing gross cardiovascular-pulmonary
function were restored to normal during
controlled ventilation, except for the four
included dogs with complications de-
scribed previously. At 24 hrs, there were
no major increases of alveolar-arterial
PO2gradients; no dog was hypoxemic.
Liver function test values were tran-
Figure 1. Tympanic membrane temperatures during exsanguination cardiac arrest of 60 mins of
no-flow in study A (A) and 90–120 mins of no-flow in study B (B). Resuscitation was with cardiopul-
monary bypass (CPB). Data are given as mean and SD.
Figure 2. Outcome in terms of final overall performance categories (OPC 1–5) at 72 hrs after
exsanguination cardiac arrest of 60 mins no-flow in study A and cardiac arrest of 90–120 mins no-flow
in study B. Each dot represents one dog. NDS, neurologic deficit scores (1–100%); HDS, total brain
histologic damage scores (0–40, no or mild damage; 40–100, moderate damage; ?100, severe
damage). *n ? 4 and †n ? 5 (one dog each was maintained for cognitive function testing).
1527Crit Care Med 2003 Vol. 31, No. 5