Evolution of the Stone Heart After

Prolonged Cardiac Arrest*

Kada Klouche, MD; Max Harry Weil, MD, PhD, Master FCCP; Shijie Sun, MD;
Wanchun Tang, MD, FCCP; Heitor P. Povoas, MD; Takashi Kamohara, MD;
and Joe Bisera, MSEE

Objectives: We hypothesized that progressive impairment in diastolic function during cardiopul-

monary resuscitation (CPR) precedes evolution of the “stone heart” after failure of CPR. We
therefore measured sequential changes in left ventricular (LV) volumes and free-wall thickness
of the heart during CPR in an experimental model.
Design: Prospective, observational animal study.
Setting: Medical research laboratory in an university-affiliated research and educational institute.
Subjects: Domestic pigs.
Methods: Ventricular fibrillation (VF) was induced in 40 anesthetized male domestic pigs
weighing between 38 kg and 43 kg. After 4 min, 7 min, or 10 min of untreated VF, electrical
defibrillation was attempted. Failing to reverse VF in each instance, precordial compression at a
rate of 80/min was begun coincident with mechanical ventilation. Coronary perfusion pressures
(CPPs) were computed from the differences in time-coincident diastolic aortic and right atrial
pressures. Left ventricular (LV) systolic and diastolic ventricular volumes and thickness of the LV
free wall were estimated with transesophageal echocardiography. The stroke volumes (SVs) were
computed from the differences in decompression diastolic and compression systolic volumes.
Free-wall thickness was measured on the hearts at autopsy.
Results: Significantly greater CPPs were generated with the 4 min of untreated cardiac arrest.
Progressive reductions in LV diastolic and SV and increases in LV free-wall thickness were
documented with increasing duration of untreated VF. A stone heart was confirmed at autopsy in
each animal that failed resuscitative efforts. Correlations with indicator dilution method and
physical measurements at autopsy corresponded closely with the echocardiographic measure-
ments.
Conclusion: Progressive impairment in diastolic function terminates in a stone heart after
prolonged intervals of cardiac arrest. (CHEST 2002; 122:1006 –1011)

Key words: cardiopulmonary resuscitation; left ventricular diastolic volume; left ventricular compliance; left ventricular
wall thickness; stone heart

Abbreviations: CPP ⫽ coronary perfusion pressure; CPR ⫽ cardiopulmonary resuscitation; LV ⫽ left ventricular;
Petco2 ⫽ end-tidal Pco2; SV ⫽ stroke volume; VF ⫽ ventricular fibrillation

C tation
urrent methods of closed-chest cardiac resusci-
lose effectiveness when the duration of
cardiac arrest prior to attempted cardiac resuscita-
tion increases to ⬎ 8 min.1,2 Accordingly, the success
*From The Institute of Critical Care Medicine (Drs. Klouche,
Weil, Povoas, and Kamohara), Palm Springs; and The Keck
School of Medicine of the University of Southern California (Drs.
Weil, Sun, Tang, and Prof. Bisera), Los Angeles, CA.
Manuscript received April 28, 2000; revision accepted March 18,
2002.
Supported in part by National Institutes of Health grant HL-
54322 from the National Heart, Lung, and Blood Institute,
Bethesda, MD, the Desert Health Care District and the Mason
Foundation, Inc.
Correspondence to: Max Harry Weil, MD, PhD, Master FCCP,
The Institute of Critical Care Medicine, 1695 North Sunrise Way,
Building 3, Palm Springs CA; e-mail: weilm@911research.org
of current methods of closed-chest compression
after protracted intervals of untreated cardiac arrest
⬎ 8 min is remote.3–5 The duration of cardiac arrest
prior to the start of cardiopulmonary resuscitation
(CPR) in human victims is the best single predictor
of outcome.6,7
In settings of regional myocardial ischemia due to
coronary artery disease, decreases in ventricular
compliance with myocardial stunning are well docu-
mented.8,9 Decreases in ventricular compliance have
also been documented during the global myocardial
ischemia of cardiac arrest.10 Reductions in left ven-
tricular (LV) end-diastolic volumes would explain, at
least in part, progressive decreases in stroke volume
(SV) during CPR. We therefore anticipated that

1006 Laboratory and Animal Investigations

decreases in LV chamber size with increases in wall
thickness would correspond to the duration of un-
treated ventricular fibrillation (VF) and terminate in
an anatomically “stony heart.” In the present study,
we validated transesophageal echocardiography in a
porcine model of cardiac arrest to quantitate the
dynamic changes in LV volumes and wall thickness
after untreated cardiac arrest for intervals of 4 min,
7 min, and 10 min and failed resuscitation. We
subsequently related these to physical measurements
at autopsy.
Materials and Methods
Animals received humane care in compliance with principles of
laboratory animal care formulated by the National Society for
Medical Research, and care and use of laboratory animals as
mandated by the Institute of Laboratory Animal Resources. The
protocols were approved by the Institutional Animal Care and
Use Committee of the Institute of Critical Care Medicine.
Animal Preparation
Male domestic pigs from a single source breeder weighing
between 38 kg and 45 kg were investigated. Animals were fasted
overnight except for free access to water. Anesthesia was initiated
by IM injection of ketamine, 20 mg/kg, followed by ear vein
injection of sodium pentobarbital, 30 mg/kg. Additional doses of
sodium pentobarbital, 8 mg/kg, were injected at intervals of
approximately 1 h to maintain anesthesia. After endotracheal
intubation, the animals received mechanical ventilation with a
tidal volume of 15 mL/kg and a peak flow of 40 L/min of room air
with the aid of a volume-controlled ventilator (Model MA-1;
Puritan Bennett; Carlsbad, CA). End-tidal Pco2 (Petco2) was
monitored with a mainstream infrared analyzer (Model 01R-
7101A; Nihon Kohden; Tokyo, Japan). Respiratory frequency was
adjusted to maintain Petco2 between 35 mm Hg and 40 mm Hg
prior to cardiac arrest without adjustment thereafter. Conven-
tional ECG scalar limb leads were continuously recorded. The
femoral artery and vein were surgically isolated under aseptic
conditions. An 8F angiographic catheter (Model 6523; USCI,
C.R. Bart; Billerica, MA) was advanced through the right femoral
artery into the descending thoracic aorta for measurement of
aortic pressure. A multilumen, thermistor and balloon-tipped
pulmonary artery catheter (41216 – 01; Abbott Critical Care;
Anaheim, CA) was flow directed from the right femoral vein into
the pulmonary artery. The atrial port was utilized for measure-
ment of atrial pressure and for injection of thermal tracer.
Positions of the catheters were guided by characteristic pressure
pulse morphology and confirmed by fluoroscopy.
For the measurements of LV volumes, a 5-MHz, single-plane,
transesophageal echocardiographic transducer with 5-MHz con-
tinuous-wave Doppler echocardiography with four-way flexure
capability (Model 21363A; Hewlett-Packard, Medical Products
Group; Andover, MA) was utilized as a noninvasive option so that
the bony chest was preserved for conventional chest compres-
sion. The probe was advanced from the incisor teeth into the
esophagus for a distance of approximately 35 cm. Feasibility
studies in our porcine model demonstrated that biplane trans-
ducers yielded less consistent images for mensuration during
precordial compression.
For inducing VF, a 4F pacing wire (EP Technologies; Moun-
tain View, CA) was advanced through a surgically isolated right
www.chestjournal.org
cephalic vein into the right ventricle. The tip of the pacing wire
was impinged on the apical endocardium during fluoroscopic
imaging. A characteristic endocardial injury current confirmed
appropriate placement.
Experimental Procedures
Baseline recordings of hemodynamic and echocardiographic
parameters were obtained. Blood temperature was maintained at
38 ⫾ 0.5°C with infrared heating lamps. VF was induced by
progressively increasing the alternating current delivered to the
endocardium from 0 to 2 mA. Mechanical ventilation was
discontinued after confirmation of VF. After 4 min, 7 min, or
10 min of untreated VF, electrical defibrillation was attempted.
Up to three 150-J biphasic electrical shocks were delivered
between the positive right infraclavicular electrode and the
negative apical electrode. If VF was not reversed after a sequence
of three shocks, precordial compression was started with a
pneumatic chest compressor (Thumper, Model 1000; MI Instru-
ments; Grand Rapids, MI) at a rate of 80 compressions per
minute. Coincident with start of precordial compression, the
animals received mechanical ventilation with a tidal volume of
15 mL/kg and a fraction of inspired oxygen of 1.0. Chest
compression was synchronized to provide a compression/ventila-
tion ratio of 5:1 with equal compression-relaxation intervals (ie, a
50% duty cycle). The compression force was adjusted to decrease
the anteroposterior diameter of the chest by 25%. After each min
of precordial compression, another sequence of up to three
shocks was delivered. Successful resuscitation was defined as a
return of a supraventricular rhythm that generated a mean aortic
pressure of ⱖ 60 mm Hg, for an interval ⬎ 5 min. When we
failed to restore spontaneous circulation after 15 min, resuscita-
tion efforts were discontinued and autopsy was performed.
Measurements
Dynamic data, including aortic pressure, right atrial pressure,
Petco2, and lead 2 of the scalar ECG, together with the digital
output of the Hewlett-Packard Acoustic Quantification system,
were recorded continuously on a personal computer-based data
acquisition system, supported by CODAS hardware and software
(Dataq Instruments; Akron, OH). Sixteen channels were pro-
vided for continuous recording at optimal sampling frequencies.
The coronary perfusion pressure (CPP) was digitally computed
from the differences in time-coincident diastolic (compression)
aortic and right atrial pressures and displayed in real-time.
Echocardiographic measurements were made frame-by-frame
on the long axis, two-chamber view. Compression systole was
defined as the minimal chamber dimension of the LV during
chest compression. Compression diastole was defined as the
maximal chamber dimension of the LV following release of chest
compression. LV areas were calculated by the method of discs
(Acoustic Quantification Technology; Hewlett Packard; Andover,
MA). LV systolic and diastolic volumes and SV were computed
using the simplified Simpson’s rule.11 LV wall thickness was
measured in the posterior wall, 1 cm distal to the mitral valve.
This site provided more consistent images in contrast to the
anterior and lateral walls.
Free-wall thickness was physically measured at autopsy in the
four animals for comparison with echocardiographic measure-
ments. It was only after analyses of the ultrasound data pointed to
the evolution of the stone heart that we performed anatomic
correlation at autopsy within 5 min of failed CPR in the final four
animals enrolled in this experimental study.
Additional studies were performed in five animals in whom VF
was induced and maintained for 7 min. The same procedure
CHEST / 122 / 3 / SEPTEMBER, 2002 1007
described earlier was used but, in addition, thermodilution
cardiac outputs were measured during CPR with the aid of a
cardiac computer (Model 9520A; American Edwards Laborato-
ries; Irvine, CA) following injection of 5 mL of saline solution
maintained at 4°C. Duplicate measurements of thermodilution
cardiac output in each instance differed by ⬍ 5% and were
averaged. These measurements of cardiac output were compared
with echocardiographic estimates of cardiac output.

Statistical Analysis

Data are presented as mean ⫾ SD. Differences in LV volumes

and free wall thickness between groups of animals were analyzed
by analysis of variance. Comparisons between time-based mea-
surements within each group were performed with analysis of
variance repeated measurements with adjustments for multiple
comparisons (Newman-Keuls). Correlations were computed by
linear regression analyses utilizing STAT VIEW II software Figure 1. CPPs during the initial 11 min of precordial compres-
(Abacus Concepts; Berkeley, CA). The outcome differences were sion and after 4 min, 7 min, or 10 min of untreated VF. The
numbers in parentheses refer to the number of unresuscitated
analyzed with the Fisher exact test and ␹2 square test. A p value
animals. BL ⫽ baseline.
⬍ 0.05 was regarded as statistically significant.

Results End-systolic volumes did not differ among the

No differences in baseline hemodynamic mea-
surements and values of LV volumes and free-wall
thickness between the three groups of animals were
observed. After 4 min of untreated VF, all five
animals obtained spontaneous circulation within
2 min of the start of CPR. As anticipated, survival
significantly decreased after 7 min and 10 min of
untreated cardiac arrest with respect to the 4-min
group. Differences were, however, not statistically
significant between 7 min and 10 min of untreated
VF. The principal outcomes are shown in Table 1.
Consistently greater CPP was documented in an-
imals after 4 min of untreated VF, as shown in
Figure 1. However, the differences between the
7-min and 10-min groups were not statistically sig-
nificant. Petco2 was greater after both 4 min and
7 min when compared to 10 min of untreated VF.
However, the differences between the 7-min and
10-min groups once again were numerically but not
statistically different (Fig 2).
three groups. The most impressive change was an
early and highly significant decrease in diastolic
volumes with increasing duration of cardiac arrest
(Fig 3). This accounted for progressive reductions in
SV (Fig 4). After 7 min of untreated cardiac arrest
and at 10 min after start of resuscitation efforts, the
diastolic volume was reduced to one third of control
values. After 10 min of untreated cardiac arrest and
10 min of CPR, the diastolic volume reduced to only
15% of baseline values. Concurrently, progressive
increases in the thickness of the free wall of the LV
were documented (Fig 5). The differences in wall
thickness between the 7-min and 10-min groups
proved to be highly significant. Autopsy demon-
strated a strikingly firm and contracted heart that,
in each instance, was characteristic of the stone
heart previously described.12,13 Ventricular cham-

Table 1—The Effects of Arrest Time on Resuscitation

and Left Ventricular Function*

Duration of Untreated VF, min

Variables 4 7 10

Resuscitated, No./total 5/5‡ 16/20 11/15

Diastolic volume, mL† 33 ⫾ 4§ 30 ⫾ 8§ 23 ⫾ 3
SVs, mL† 25 ⫾ 3§ 21 ⫾ 7§ 14 ⫾ 2
Free-wall thickness, mm† 4.8 ⫾ 0.6§ 5.0 ⫾ 0.6§ 6.5 ⫾ 0.5
Figure 2. Petco2 during the initial 11 min of precordial
*Data are presented as mean ⫾ SD unless otherwise indicated. compression and after 4 min, 7 min, or 10 min of untreated VF.
†Measurements obtained 1 min after start of precordial compression. The number of unresuscitated animals included in each time
‡p ⬍ 0.05 vs 7 and 10 min. interval is shown in Figure 1, as is the expansion of abbreviation.
§p ⬍ 0.01 vs 10 min.

1008 Laboratory and Animal Investigations

Figure 3. Diastolic LV volumes during the initial 11 min of
precordial compression and after 4 min, 7 min, or 10 min of
untreated VF. The number of unresuscitated animals included in
each time interval is shown in Figure 1. ML ⫽ milliliter; see
Figure 1 for expansion of other abbreviation.
ber dimensions were markedly decreased, and the
thickness of the septum and the free wall were
markedly increased. The correlation between echo-
cardiographic measurements and physical measure-
ment of free-wall thickness on four animals that
failed resuscitation attempts after 7 min of untreated
cardiac arrest are shown in Figure 6. In the five
animals in which echocardiographic estimates of SV
were compared with thermodilution measurements
during CPR, the correlation (r2) was 0.94,
p ⬍ 0.0001 (Fig 7).
Discussion
When immediate defibrillation fails, the success of
CPR becomes contingent on prompt restoration of
Figure 4. SVs during the initial 11 min of precordial compres-
sion and after 4 min, 7 min, or 10 min of untreated VF. The
number of unresuscitated animals included in each time interval
is shown in Figure 1. See Figures 1 and 3 for expansions of
abbreviations.
www.chestjournal.org
Figure 5. LV free-wall thickness during the initial 11 min of
precordial compression and after 4 min, 7 min, or 10 min of
untreated VF. The number of unresuscitated animals included in
each time interval is shown in Figure 1, as is the expansion of
abbreviation.
CPP and therefore myocardial blood flow.14,15
Closed-chest compression provides sufficient for-
ward blood flow in the early intervals following
cardiac arrest. However, when CPR is started after
7 min or 10 min of untreated cardiac arrest, as in the
present investigations, precordial compression typi-
cally fails to generate sufficient forward blood flow
and perfusion pressures to be predictive of success-
ful resuscitation.1,5,15,16 Progressive decreases in LV
diastolic volumes and increases in LV wall thickness
characterize the diastolic dysfunction that is associ-
ated with decreased SV. The findings are consistent
with a progressive loss of LV compliance associated
with ischemic contracture following global myocar-
dial ischemia as previously reported by our group.17
After VF is induced in the isolated beating heart,
progressive increases in LV diastolic pressure during
VF were documented.18 Whether compression of
Figure 6. The relationship between echocardiographic and
physical measurements at autopsy of LV free-wall thickness on
four animals.
CHEST / 122 / 3 / SEPTEMBER, 2002 1009

Figure 7. The relationship between echocardiographic and
thermodilution measurements of SVs during CPR on five
animals.
the chest with increases in intrathoracic pressure also
accounts for decreases in preload and therefore SV is
postulated and such may be potentially minimized by
active compression decompression CPR.19
Our findings of ischemic contracture are reminis-
cent of observations reported in settings of regional
ischemia.8,9,20,21 The concept of ischemic contrac-
ture of the myocardium is associated with myocardial
stunning. The same concept may be applicable to the
global myocardial ischemia of cardiac arrest. Visner
et al10 compressed the left main coronary artery and
produced decreases in LV compliance, much like
those observed by us in the settings of failed cardiac
resuscitation. Augmented crossbridging between ac-
tin and myosin followed depletion of high-energy
phosphates.13,22 Neumar et al23,24 found a strong
temporal correlation between the duration of global
myocardial ischemia of cardiac arrest and high-
energy phosphates. These time-dependent changes
were associated with reductions in ventricular dia-
stolic volumes and increases in myocardial stiffness.
These, in turn, accounted for decreased effective-
ness of chest compression for producing forward
blood flow during CPR.25 Increases in ventricular
wall stiffness of themselves account for reductions in
diastolic and SVs because of ischemia consequent to
increased mechanical resistance to coronary blood
flow.26 Accordingly, both ischemic stunning and
increased coronary arterial resistance to blood flow
are implicated. Histopathologic examinations of
stone hearts following prolonged cardiac arrest in
human victims have demonstrated myocardial
edema with increased myocardial tissue water as an
additional feature.27,28 Ultrastructural studies on
hearts following cardiac resuscitation are eagerly
awaited to explain both the hemodynamic and gross
anatomic abnormalities herein described.
We acknowledge potential limitations in the inter-
1010
pretation of our findings. Pentobarbital, which is well
established as an anesthetic in CPR settings, is a
cardiac depressant and a coronary dilator.29 –33 Be-
cause all animals in the present study received the
same doses by weight, which are less than those
recognized as cardiodepressant, the differences be-
tween groups are not likely to reflect the effects of
the anesthetic agent.34
LV volume, free-wall thickness, and dynamic
changes associated with progressive impairment of
diastolic function are consistent with decreased com-
pliance. However, we elected to forego measure-
ments of LV pressures to minimize previously ob-
served obfuscating effects of intracardiac catheters
for computation of ventricular compliance. The de-
creases in LV chamber volumes in association with
increases in wall thickness provide us with secure
evidence of a contracted ventricle. A monoplane
transducer comparable to that employed by earlier
investigators was utilized to obviate major artifacts
induced by chest compression.35,36 The single-plane,
area-length method of measurement was utilized
with the assumption that the LV is best represented
by a conical ellipsoid.11 However, LV volumes when
measured in the transesophageal long-axis view,
yield comparable values in pigs to those measured
through a transgastric window.37,38 Nevertheless, we
have independently validated this method under the
conditions of the present study.
Finally, we caution against direct extrapolation of the
present findings to clinical CPR. The porcine model
was chosen because the porcine thorax is comparable
to that of humans, and this model has been extensively
exercised and validated against human observations
during CPR.29 –31,39 Nevertheless, the animals were
young, healthy pigs, free of atherosclerotic disease. It is
within the context of these limitations that the present
study traces the evolution of the stone heart.
References
1 Lee SK, Vaagenes P, Safar P, et al. Effect of cardiac arrest
time on cortical cerebral blood flow during subsequent
standard external cardiopulmonary resuscitation in rabbits.
Resuscitation 1989; 17:105–117
2 Gazmuri RJ, Weil MH, von Planta M, et al. Cardiac resusci-
tation by extra corporeal circulation after failure of conven-
tional CPR. J Lab Clin Med 1991; 101:984 –988
3 Eisenberg MS, Hallstrom AP, Copass MK, et al. Treatment of
ventricular fibrillation: emergency medical technician defi-
brillation and paramedics services. JAMA 1984; 251:1723–
1726
4 Stueven H, Troiano P, Thompson B, et al. Bystander/first
responder CPR: ten years experience in a paramedic system.
Ann Emerg Med 1986; 15:707–710
5 Duggal C, Weil MH, Tang W, et al. Effect of arrest time on
the hemodynamic efficacy of precordial compression. Crit
Care Med 1995; 23:1233–1236
Laboratory and Animal Investigations
 6 Bedell SE, Delbanco TL, Cook F, et al. Survival after
cardiopulmonary resuscitation in the hospital. N Engl J Med
1983; 309:569 –576
7 Copley DP, Mantle JP, Rodgers WJ. Improved outcome for
prehospital cardiopulmonary collapse with resuscitation by
bystanders. Circulation 1977; 56:901–905
8 Braunwald E, Kloner RA. The stunned myocardium: pro-
longed, post-ischemic ventricular dysfunction. Circulation
1982; 66:1146 –1149
9 Charlat ML, O’Neill PG, Hartley CJ, et al. Prolonged abnor-
malities of left ventricular diastolic wall thinning in the
“stunned” myocardium in conscious dogs: time course and
relation to systolic function. J Am Coll Cardiol 1989; 13:
185–194
10 Visner MS, Arentzen CE, Parrish DG, et al. Effects of global
ischemia on the diastolic properties of the left ventricle in the
conscious dog. Circulation 1985; 71:610 – 619
11 American Society of Echocardiography Committee on Stan-
dards. Recommendations for quantification of the left ventri-
cle by two-dimensional echocardiography. J Am Soc Echo-
cardiogr 1989; 2:358 –367
12 Takino M, Okada Y. Firm myocardium in cardiopulmonary
resuscitation. Resuscitation 1996; 33:101–106
13 Kingsley-Hickman PB, Sako EY, Yang MQ, et al. Ischemic
contracture begins when anaerobic glycolysis stops: a 31P-
NMR study of isolated rat hearts. Am J Physiol 1991;
261:H469 –H478
14 Ditchey RV, Winkler JV, Rhodes CA. Relative lack of coro-
nary blood flow during closed-chest resuscitation in dogs.
Circulation 1982; 66:297–302
15 Niemann JT, Criley JM, Rosborough JP, et al. Predictive
indices of successful cardiac resuscitation after prolonged
arrest and experimental cardiopulmonary resuscitation. Ann
Emerg Med 1985; 14:521–528
16 Brown CG, Katz SE, Werman HA, et al. The effect of
epinephrine vs methoxamine on regional myocardial blood
flow and defibrillation rates following a prolonged cardiore-
spiratory arrest in swine model. Am J Emerg Med 1987;
5:362–369
17 Gazmuri RJ, Weil MH, Tang W, et al. Progressive loss of
ventricular compliance during closed-chest cardiac resuscita-
tion [abstract]. Circulation 1994; 90:I287
18 Gazmuri RJ, Berkowitz M, Cajigas H. Myocardial effects of
ventricular fibrillation in the isolated rat heart. Crit Care Med
1999; 27:1542–1550
19 Tucker KJ, Redberg RF, Schiller NB, et al. Active compres-
sion-decompression resuscitation: analysis of transmitral flow
and left ventricular volume by transesophageal in humans.
J Am Coll Cardiol 1993; 22:1485–1493
20 Ross J. Is there a true increase in myocardial stiffness with
acute ischemia? Am J Cardiol 1989; 63:87E–91E
21 Palacios I, Johnson RA, Newell JB, et al. Left ventricular
end-diastolic pressure volume relationships with experimen-
tal acute global ischemia. Circulation 1976; 3:428 – 436
22 Johnson BA, Weil MH. Redefining ischemia due to circula-
tory failure as dual defects of oxygen deficits and of carbon
dioxide excesses. Crit Care Med 1991; 11:1432–1438
www.chestjournal.org
23 Neumar RW, Brown CG, Ligten PV, et al. Estimation of
myocardial ischemic injury during ventricular fibrillation with
total circulatory arrest using high-energy phosphates and
lactate as metabolic markers. Ann Emerg Med 1991; 20:
222–229
24 Neumar RW, Brown CG, Robitaille PML, et al. Myocardial
high energy phosphate metabolism during ventricular fibril-
lation with total circulatory arrest. Resuscitation 1990; 19:
199 –226
25 Sharff JA, Pantley G, Noel E. Effect of time on regional organ
perfusion during two methods of cardiopulmonary resuscita-
tion. Ann Emerg Med 1984; 13:649 – 656
26 Hottenrott C, Maloney JV, Buckberg G. Studies of the effects
of ventricular fibrillation on the adequacy of regional myo-
cardial flow, III: mechanisms of ischemia. Thorac Cardiovasc
Surg 1974; 68:634 – 645
27 Tabayashi K, McKeown PP, Miyamoto M, et al. Diastolic
mechanical properties of the left ventricle after global isch-
emia. J Cardiovasc Surg 1991; 32:250 –258
28 Goto R, Tearle H, Steward DJ, et al. Myocardial oedema and
ventricular function after cardioplegia with added mannitol.
Can J Anaesth 1991; 38:7–14
29 Tang W, Weil MH, Noc M, et al. Augmented efficacy of
external CPR by intermittent occlusion of the ascending
aorta. Circulation 1993; 88:1916 –1921
30 Tang W, Weil MH, Sun S, et al. Epinephrine increases the
severity of postresuscitation myocardial dysfunction. Circula-
tion 1995; 92:3089 –3093
31 Tang W, Weil MH, Sun S, et al. The effects of biphasic and
conventional monophasic defibrillation on postresuscitation
myocardial function. J Am Coll Cardiol 1999; 34:815– 822
32 von Planta I, Weil MH, von Planta M, et al. Cardiopulmonary
resuscitation in the rat. J Appl Physiol 1988; 65:2641–2647
33 Gazmuri RJ, Weil MH, Bisera J, et al. Myocardial dysfunction
after successful resuscitation from cardiac arrest. Crit Care
Med 1996; 24:992–1000
34 Urruh HW, Wang R, Bose D, et al. Does pentobarbital
anesthesia depress left ventricular contractility in dogs? Am J
Physiol 1991; 261:H700 –H706
35 Varriale P, Maldonado JM. Echocardiographic observations
during inhospital cardiopulmonary resuscitation. Crit Care
Med 1997; 10:1717–1720
36 van der Wouw PA, Koster RW, Delemarre BJ, et al. Diag-
nostic accuracy of transesophageal echocardiography during
cardiopulmonary resuscitation. J Am Coll Cardiol 1997; 30:
780 –783
37 Porter TR, Ornato JP, Guard CS, et al. Transesophageal
echocardiography to assess mitral valve function and flow
during cardiopulmonary resuscitation. Am J Cardiol 1992;
70:1056 –1060
38 Smith MD, MacPhail B, Harrison MR, et al. Value and
limitations of transesophageal echocardiography in determi-
nation of left ventricular volumes and ejection fraction. J Am
Coll Cardiol 1992; 19:1213–1222
39 Gudipati C, Weil MH, Bisera J, et al. Expired carbon dioxide:
a noninvasive monitor of cardiopulmonary resuscitation. Cir-
culation 1988; 77:234 –239
CHEST / 122 / 3 / SEPTEMBER, 2002 1011