imaging of the apparent diffusion coefficient for the evaluation of cerebral metabolic recovery...

Pergamon

l Original Contribution

Magnetic Resonance Imaging, Vol. 13. No. 6, pp. 781-790, 1995 Copyright 0 1995 Elsevier Science Inc. Printed in the USA. All rights reserved

0730-725X/95 $9.50 + .OO

0730-725X(95)00030-5

IMAGING OF THE APPARENT DIFFUSION COEFFICIENT FOR THE EVALUATION OF CEREBRAL METABOLIC

RECOVERY AFTER CARDIAC ARREST

MATTHIAS FISCHER, * KURT BOCKHORST, MATHIAS HOEHN-BERLAGE, BERND SCHMITZ, * AND KONSTANTIN-ALEXANDER HOSSMANN

Max-Planck-Institute for Neurological Research, Department of Experimental Neurology, Cologne, Germany

The apparent diffusion coefficient (ADC) of water is a sensitive indicator of water and ion homeostasis of brain. Resuscitation of the brain after cardiac arrest, the most frequent reason for global cerebral ischemia under clinical conditions, depends critically on the reversal of disturbances of water and ion homeostasis. We, therefore, investi- gated whether ADC imaging can be used to monitor the development and reversal of ischemic brain injury during and after cardiac arrest. Ten adult mongrel normothermic cats were anesthetized with alfentanil and midazolam, immobilized with pancuronium, and mechanically ventilated with Oz/N20. Arterial, left ventricular, central venous, and intracranial pressures were monitored throughout the experiment. Magnetic resonance imaging was performed in a 4.7T MR scanner with a shielded gradient system. Diffusion-weighted images (DWI) were obtained by pulsed gradient spin echoes (diffusion weighting factor b: 0,500, 1000, 1500 s/mm’). Quantitative ADC images were calculated from DWIs by fitting signal intensities against b-factors. Fifteen minute cardiac arrest was induced in the magnet by electrical fibrillation. Resuscitation was also carried out in the magnet, using a pneumatic vest for remotely controlled closed chest cardiac massage. Seven of 10 animals were resuscitated successfully and subsequently monitored for 3 h. During cardiac arrest, ADC declined from 678 f 79 x 10m6 to 430 f 128 x 10m6 mm*/s (63% of baseline). In the successfully resuscitated animals ADC returned to 648 2 108 x 10m6 mm*/s within 30 min and remained at this level throughout the 3 h of recirculation. Regional evaluations of ADC revealed a transient overshoot in brainstem and basal ganglia to 114% of control at 15 min before return- ing to baseline values after 40 min. Failure of cardiac resuscitation prevented ADC normalization and led to its further decline to below 50% of control. Postcardiac arrest normalization of ADC maps correlated with homogeneous return of ATP, glucose, and lactate to near normal, whereas failure of ADC normalization was associated with depletion of ATP and glucose and severe lactate accumulation. In conclusion, our data indicate, that normalization of ADC is a reliable indicator of cerebral recovery after resuscitation from cardiac arrest.

Keywords: Diffusion-weighted NMR imaging; ADC; Cerebral metabolism; Cardiac arrest; Resuscitation; Pneu- matic vest.

INTRODUCTION

In the normothermic mammal, complete interruption of cerebral blood flow induced by either cardiac arrest or vascular occlusion results in the rapid breakdown of energy metabolism and the inhibition of energy-

dependent ion exchange pumps.’ This leads to the depolarization of cell membranes, an equilibration of transmembrane ion gradients, and a massive shift of fluid from the extracellular into the intracellular com- partment .2*3 Restoration of blood flow and energy metabolism after ischemia reverses these changes: ion

RECEIVED 10/21/94; ACCEPTED 4/24/95. Address correspondence to Professor Dr. K.-A. Hoss-

mann, Max-Planck-Institut fur neurologische Forschung, Abteilung fur experimentelle Neurologie, Gleueler Strage 50,

D-5093 1 Koln, Germany. *On leave from Department of Anesthesiology and Crit-

ical Care Medicine, University of Bonn, Bonn, Germany.

781

782 Magnetic Resonance Imaging 0 Volume 13, Number 6, 1995

pumps are reactivated, and water homeostasis is restored.3 Water compartmentation is, therefore, a sensitive marker of ischemic injury and its reversal.

With the advent of new NMR imaging modalities, which allow mapping of the apparent diffusion coefficient (ADC) of water, ischemic changes of water compartmentation can be detected noninvasively with high spatial and temporal resolution.495 Although the pre- cise mechanisms of ischemic alterations of brain water diffusion are not known, there is compelling evidence that the changes of the ADC correlate closely with the changes of extracellular fluid volume.6y7 Imaging of water diffusion, therefore, provides an indirect mea- surement of brain energy state.

In a previous study, ADC imaging was used to monitor brain injury and its reversal after 1 h complete cerebrocirculatory arrest in normothermic cats, induced by reversible occlusion of the innominate and subclavian arteries.’ Ischemia produced a stereotyped homogeneous decline of ADC, but the recovery differed greatly, depending on the general hemodynamic state. In animals with reperfusion of the brain starting at low blood pressure, ADC recovery was heterogeneous or absent, whereas in animals with high reperfusion pressure, ADC returned to normal. Comparison of ADC with bioluminescence images of energy state confirmed the close correlation of the recovery of ADC with energy metabolism.’

The importance of the general hemodynamic situa- tion for postischemic brain resuscitation is also re- flected by the results reported following cardiac arrest, which is the most frequent reason for global cerebral ischemia under clinical conditions. Using conventional extrathoracal cardiac massage, brain energy metabolism and electrophysiological function could be consistently restored only after cardiac arrest of up to 15 min,s but recovery occasionally succeeded even after cardiac arrest of as long as 30 min, when reperfusion pressure was improved by two-stage resuscitation,8 open chest cardiac massage,‘or by extracorporeal bypass perfusion.” The previously reported severe neurological deficits after lo-12 min cardiac arrest11-‘4 may, therefore, be due to early postischemic hemodynamic dis- turbances15 or delayed postischemic hypoperfusion.‘6 However, positive evidence is lacking, because of the impossibility to monitor, with adequate temporal and spatial resolution, the regional pattern of blood flow and metabolism throughout the postischemic resuscitation process.

To improve resuscitation of the brain following cardiac arrest and to develop new therapeutical strategies, this difficulty has to be overcome. ADC mapping for the noninvasive regional evaluation of the postischemic recovery process may be a valuable tool. However,

application of this methodology under experimental conditions requires remote induction and reversal of cardiac arrest inside the magnet. In the present investigation, we report such an experiment and describe, for the first time, the full temporal evolution of regional ADC changes during and after 15-min cardiac arrest in cats as well as the relationship of ADC to regional metabolic alterations.

MATERIALS AND METHODS

Animal Preparation Animal experiments were carried out according to

the Society for Neurosciences (SFN) guidelines, and were approved by the ethical committee of the local authorities.

Ten adult male mongrel cats (2.6-4.7 kg body weight) were anesthetized with 10 mg/kg ketamine, followed by inhalation anesthesia with 1.5% halothane in 70/30% NzO/Oz. After endotracheal intubation animals were immobilized with pancuronium bromide (0.12 mg/kg/h) and mechanically ventilated with a tidal volume of 15 ml/kg, using a small animal ventilator (Harvard Apparatus, South Natick, MA). Ventilation rate was adjusted to maintain an arterial carbon diox- ide pressure (PaC02) of 30-35 mmHg.

The left femoral vein was cannulated with a 5-French triple-lumen catheter for drug administration and monitoring of the central venous pressure (Kom- bidyn Monitoring Set, Braun Melsungen, Germany). Both femoral arteries were cannulated with fluid-filled 3-French catheters to measure the aortic and the left ventricular pressure (Kombidyn Monitoring Set, Braun, Melsungen, Germany). A PE-50 catheter was inserted into the bladder to monitor urine production. Rectal temperature was measured with a thermistor probe and kept constant at 37°C using a feed-back controlled heated water jacket wrapped around the body of the animal.

The head of the animal was mounted in a stereotaxic frame, the skin over the apex of the skull was removed, and a small cranial window was prepared over the parietal region. A Plexiglas tube of 5 mm diameter was cemented watertight to the window, leaving the dura intact. Intracranial pressure (ICP) was measured using a pressure transducer (Kombidyn Monitoring Set, Braun, Melsungen, Germany) connected to the tube by a fluid-filled catheter. Two other burr holes were prepared over the parietal region to insert nonmagnetic graphite electrodes for recording of the electroencephalogram (EEG).

At the end of the surgical preparation, halothane was gradually substituted by intravenous infusion of alfentanil(0.3 mg/kg/h) and midazolam (0.75 mg/kg/h)

ADC recovery after cardiac arrest 0 M. FIXHER ET AL. 783

to avoid adverse side effects of halothane on cardiac function. Fluid homeostasis was preserved by infus- ing Ringer’s solution at 4 ml/kg/h throughout the experiment.

Induction of 15-Minute Cardiac Arrest and Resuscitation Protocol

Induction of cardiac arrest, closed-chest cardiac massage, defibrillation, and 180 min reperfusion after successful resuscitation were carried out in the magnet of the tomographic system by remote control. Cardiac arrest was induced by electrical fibrillation for 45 s (AC 30-60 V, 5 mA) through flat, flexible copper defibrillator electrodes attached to the chest of the animal. Fol- lowing ventricular fibrillation, mechanical ventilation, anesthesia, and infusions were interrupted and the heating system was switched off.

Closed chest cardiac massage was performed with a pneumatic vest developed in our laboratory. The animal was mounted in prone position in a stereotaxic frame. An inflatable rubber bag of 10 x 5 x 5 cm size (length x width x height during maximal inflation) was placed presternally and covered about 25% of thoracic circumference of the animal. The bag was connected to a tube of 3 m length with an inner diameter of 10 mm, which led to the pressure-controlling system placed out- side the magnet. A nylon cuff reinforced by a semicir- cular Plexiglas cylinder was fixed around the rubber bag and the thorax to maximize sternal displacement. Compared to the vest used by Halperin et al. ,” our de- vice produced less circumferential compression and re- sulted in excellent resuscitability.

The vest was inflated with pressurized nitrogen, using two electromagnetic valves, which controlled the nitrogen flow from the reservoir to the rubber bag and from the bag to the ambient atmosphere. The valves were operated by electronic switches with a compression rate of 120/min and a duty cycle of 30% to achieve near complete drain of the bag via the outlet valve. The power of compression was adjusted by varying the pressure in the nitrogen reservoir to generate a diastolic blood pressure of about 50 mmHg.

Cardiopulmonary resuscitation was started 15 min after onset of cardiac arrest according to the Ameri- can Heart Association (AHA) guidelines.18 First, a bo- lus of 0.2 mg/kg epinephrine was injected, followed by closed-chest cardiac massage as described above. At this time the heating system was switched on and ventilation was resumed with 100% oxygen at a rate of 140% of the prearrest frequency. Following 4 min of closed-chest cardiac massage, the first defibrillation was carried out using the external defibrillator electrodes. If this failed, an additional injection of 0.1 mg/kg epinephrine was given and external heart mas-

sage was continued for another 4 min. If ventricular fibrillation or ventricular tachycardia occurred after return of spontaneous heart function, further DC shocks were applied as necessary. All shocks were given without interrupting chest compression.

With the beginning of closed-chest cardiac massage an infusion of sodium bicarbonate (2 mval/kg/30 min) was started. Subsequently, the following drugs and so- lutions were applied: Ringer’s solution (4 ml/kg/h) to compensate for fluid loss by respiration and urine production, additional NaHC03 as required to adjust arterial base excess to normal, dopamine (3-5 pg/kg/min) to improve renal blood flow and epinephrine (0.05- 0.2 pg/kg/min) to stabilize mean arterial blood pressure above 90 mmHg.

Recording of Physiological Variables Electroencephalogram (EEG), electrocardiogram

(ECG), aortal, left ventricular, central venous, and intracranial pressures were recorded on an eight-channel polygraph (Gould Oscillograph Recorder 3000, Gould Inc., Cleveland, OH). These data were acquired, and analysis of EEG was carried out using a Macintosh II f computer (Apple, Cupertino, CA), an analog/digital converter (Macadios, GW Instruments, Somerville, USA) and the appropriate software (Superscope, Macadios). Data were acquired with a sampling rate of 125 Hz, processed and transferred to a database every 30 s. Mean arterial, central venous and intracranial pressures were averaged, and the left ventricular and aorta1 diastolic pressures were estimated from the min- imum values of each trace. Cerebral perfusion pressure was calculated from the difference of mean arterial blood pressure and the higher of either intracranial or central venous pressure. Myocardial perfusion pressure was calculated from the difference between diastolic aorta1 and diastolic left ventricular pressure.

Arterial blood samples were taken repeatedly throughout the experiment, initially at 30 min and later at 60 min intervals, for assessment of blood gases and electrolytes (Blood Gas System 288, Ciba Corning Di- agnostics, Fernwald, Germany), glucose (Glucose An- alyzer 2, Beckman Instruments, Fullerton, CA), lactate (Model 27, YSI Inc., Yellow Springs, OH), and osmolality (Vapour Pressure Osmometer, Wescor Inc., Lo- gan, UT).

NMR Methods Magnetic resonance imaging (MRI) was performed

in a BIOSPEC 4.7 T, 30 cm horizontal bore system (Bruker, Karlsruhe, Germany) equipped with actively shielded gradient coils (Bruker; maximum gradient strength per channel: 100 mT/m; rise time c 250 ps). A birdcage ‘H resonator with an inner diameter of


15.5 cm was used for excitation and reception of the RF signals. The head of the animal was fixed in a nonmagnetic head holder and positioned in the magnet. Multislice sagittal pilot scans (TR/TE 1000 ms/34 ms) were acquired for the selection of a set of seven coronal slices at Horsley-Clark coordinates A20-PlO with an interslice distance of 1.5 mm. Diffusion-weighted images (DWI) were obtained using a multislice Stejskal- Tanner-type pulsed gradient spin echo method (PGSE, TR/TE 2000 ms/34 ms, two averages, 4 min total scan time). The slice thickness was 3 mm, the field of view was 10 cm with an image matrix of 64 x 128 pixel. For quantitative determination of ADC, DWIs with different diffusion-weighting factors b (b = 0, 500, 1000, 1500 s/mm’) were recorded before and after 1, 2, and 3 h of recirculation. To improve temporal resolution during the critical phase of cardiac arrest and resuscitation, additional single DWIs (b = 1500 s/mm2) were recorded every 5 min. These and unweighted spin echoes (b = 0 s/mm2) of the previously recorded four- point measurements were used for an estimation of ADC. The resonator was disconnected from the pre- amplifier during cardiopulmonary resuscitation to avoid artefacts caused by electrical defibrillation, but image acquisition was resumed immediately after successful resuscitation, resulting in a first ADC image at 5 min after restoration of spontaneous circulation.

Computation of quantitative ADC images was performed on a VAX 3200 workstation (Digital Equipment Corporation, Maynard, MA), using a single-exponential model.19 ADC images were transferred to a Macintosh computer (Apple, Cupertino, CA). The image process- ing program IMAGE (National Institutes of Health, Bethesda, MD) was used to display sets of coronal ADC images and to calculate regional ADC changes. The mean value of ADC was calculated by averaging the ADC over all seven slices. Regional evaluations of ADC were carried out in three regions of interest, i.e., neocortex, basal ganglia, and brainstem.

Bioluminescence Imaging Three hours after resuscitation from cardiac arrest

the animals were removed from the magnet and the heads were frozen in situ with liquid nitrogen for bio- chemical analysis. 2o Frozen brains were sawed in the cold room into four coronal slices, and subsequently cut at -20°C into 20 pm cryostat sections in the same planes as the NMR images. Adjacent sections were processed for bioluminescence imaging of ATP, glucose, and lactate.2’,22

Bioluminescence images were digitized with a rotat- ing densitometer and processed with the same image analysis system as used for the display of ADC maps (see above). Bioluminescence images were quantified

by calibrating optical densities with the metabolite concentrations measured by standard enzymatic techniques in small tissue samples taken from various regions of the cryostat block.21,22

Statistical Analysis Data are expressed as means f SD. Differences were

analyzed for significance by analysis of variance and Student’s t-test. Statistical significance was assumed for p < 0.05.

RESULTS

Physiological Observations Before induction of ventricular fibrillation, all phys-

iological variables were within the normal range (Ta- ble 1). Transthoracic electrical fibrillation led to cardiac arrest within less than 20 s in all 10 cats. After 15 min cardiac arrest, resuscitation was started in the magnet, using the remotely controlled closed-chest compression system. In two animals, cardiac resuscitation did not succeed due to technical problems of the pneumatic vest, but in the other eight animals spontaneous cardiac funtion returned and arterial blood pressure increased above 120 mmHg within 8 min (mean 5.5 + 1.9 min). In four of these animals defibrillation succeeded with the first countershock applied after 4-min chest compression. Two cats needed repetitive defibrillation, and two others a second dose of 0.1 mg/kg epinephrine to reverse ventricular fibrillation. One animal died 20 min after initial successful resuscitation due to cardiogenic sho,ck and electromechanical dissociation. In the animals with successful cardiac resuscitation, mean myocardial perfusion pressure was 23 -I 7 mmHg during closed-chest cardiac massage, mean cerebral perfusion pressure was 27 + 9 mmHg, and mean arterial pressure was 59 k 12 mmHg (Table 1).

After return of cardiac function, mean arterial blood pressure transiently rose up to 180 mmHg due to the epinephrine administration at the beginning of cardiopulmonary resuscitation. Twenty minutes later mean arterial pressure returned to normal and was subsequently stabilized at about 115 mmHg.

Intracranial pressure increased to about 24 mmHg shortly after return of cardiac function and then declined to normal within 50 min in four out of seven animals. The other three cats exhibited a secondary increase to 24 + 8 mmHg at the end of the 3 h recirculation period. The cerebral perfusion pressure paral- leled the changes in blood pressure (Fig. 1). It was highest shortly after resuscitation but subsequently declined and stabilized at the prearrest value between 30 and 60 min of recirculation.

ADC recovery after cardiac arrest 0 M. FISCHER ET AL.

Table 1. Physiological variables

785

Recirculation

Control Resuscitation 10 min 60 min 120 min 180 min

MABP (mmHg) 117 k 19 59 * 12* 180 f 64* 114 f 20 116 t 24 121 f 30 CVP (mmHg) 7.9 * 1.4 33.7 + 12.1* 11.4 -+ 3.3* 8.9 f 2.0 8.7 f 2.1 6.5 zk 3.7 ICP (mmHg) 11.5 f 3.6 23.5 f 5.1* 24.3 + 10.4* 9.2 + 6.1 12.2 + 9.3 15.3 f 11.8 MPP (mmHg) 9s f 11 23 + 7* 135 f 55 ‘70 f 11 88 + 20 100 f 27 CPP (mmHg) 10s * 22 27 + 9* 158 t 58* 103 + 19 102 + 20 106 k 30

PaO, (mmHg) 175 f 36 n.d. 309 +- 70* 348 + 90* 217 + 27 203 I!z 39 PaCOz (mmHg) 3S +6 n.d. 27 +5 34 *7 30 k8 32 +4 wH 7.34 f 0.10 n.d. 7.21 -t 0.14* 7.09 It 0.13* 7.24 + O.lO* 7.27 + 0.1 Hct (%) 39 * 5 n.d. 43 -t 6* 44 f 6* 46 k5* 47 f 8* Glucose (mg/dl) 124 * 17 n.d. 270 zk 155 195 + 69* 185 f 44* 153 + 37* Lactate (mM) 1.0 f 0.2 n.d. 8.0 + 3.3* 8.8 + 4.8* 6.6 k 4.3* 4.5 + 2.6% Sodium (mM) 157 + 5 n.d. 160 -t7 161 +6 159 +9 161 k7 Potassium (mM) 3.7 -t 0.5 n.d. 6.1 + 2.0* 4.2 + 1.1 4.4 f 1.7 4.4 i 1.8 Calcium (mM) L.2 f 0.2 n.d. 1.5 + 0.6 1.1 + 0.2 1.1 f 0.1 1.1 + 0.1 Osmolality (mosm) 297 f 15 n.d. 317 z!z 19* 311 + 10* 308 + 12 315 +8*

MABP: mean arterial blood pressure, CVP: central venous pressure, ICP: intracranial pressure, MPP: myocardial perfusion pressure, CPP: cerebral oerfusion uressure. Hct: hematocrit. n.d.: not determined. All measurements (means + SD) were performed in arterial blood samples withdrawn at the indicated times. *Significantly different from control recordings (p < 0.05).

ADC Measurements The control mean value of ADC calculated by four-

point fit was 678 + 79 x 10M6 mm*/s, and by two- point fit 641 + 91 x lop6 mm2/s. During cardiac arrest, mean ADC dropped to 80%) 70070, and 63% of baseline (430 + 128 x 1O-6 mm*/s) at 5, 10, and 15 min (end of scan time), respectively. Thirty minutes after successful resuscitation mean ADC returned to 648 + 108 x 10e6 mm*/s and could be stabilized at this level throughout the 3-h recirculation period (Figs. 2 and 4). In the cat, which died due to cardiogenic shock 20 min after resuscitation, mean ADC declined to 368 x lop6 mm2/s (47% of control) within 2 h after cardiac failure, i.e., distinctly below the value at the end of 15min cardiac arrest (Fig. 3).

Regional evaluation of ADC revealed different time courses in different brain regions (Fig. 4). In the basal ganglia, ADC increased transiently to 114% of control at 15 min, before it returned to control value after 40 min. In contrast, ADC of neocortex returned to control within 15 min without such overshoot. In the brainstem, ADC dropped to a significantly lower level during cardiac arrest (5 1% of control value, p = 0.012 as compared to ADC in neocortex), followed by an overshoot to 114% of control before it returned to baseline within the first hour of recirculation.

Bioluminescence Imaging Bioluminescence images of coronal cryostat sections

were prepared at 3 h after cardiac arrest and compared

with the ADC images. In animals with normalization of ADC, lactate level was low and the content of ATP and glucose revealed a homogeneous distribution (Fig. 5). In contrast, in the animal without ADC recovery, ATP and glucose were depleted, and lactate was severely increased in all tissue sections. Quantitative enzymatic measurements in the frontal pole revealed the following concentrations: in the successfully resuscitated animals ATP was 1.72 + 0.49 pmol/g, glucose was 10.7 * 7.2 pmol, and lactate was 1.74 + 0.74 pmol/g. In the animal that died 20 min after resuscitation, ATP was 0.27 pmol/g, glucose was 0.28 pmol/g, and lactate was 19.3 fimol/g.

DISCUSSION

In the present study of 15-min cardiac arrest, mean ADC of brain water declined to 63% of control. This value is comparable to previously reported data. Hoehn- Berlage et al. 23 observed a decline to about 66% after 18 min, and Davis et a1.24 to 62% after 10 min ischemia. With longer duration of circulatory arrest, ADC does not decrease much further. In fact, in our previous study of 1 h isolated cerebrocirculatory arrest, ADC was 68% after 10 min and 63% after 50 min ischemia.’ The decrease of ADC to below 50% in the cat that died after resuscitation was, therefore, most likely due to a drop in body temperature. According to Hase- gawa et al., 25 ADC declines by 13 x lO-‘j mm*/s per degree Centigrade. The difference between ADC val-

Magnetic Resonance Imaging 0 Volume 13, Number 6, 1995

250

200

zl

g 150 E.

ii 100

%

50

0

250

200

9 150

1 g 100 0

50

0

CA

i

MEAN ARTERIAL BLOOD PRESSURE

co 0 30 60 90 120 CPR RECIRCULATION

CEREBRAL PERFUSION PRESSURE

150 180min

co 0 30 60 90 120 150 18Omin

CPR REClRClJiATlON

Fig. 1. Blood pressure and cerebral perfusion pressure of successfully resuscitated animals. CA: cardiac arrest; CPR: cardiopulmonary resuscitation. *Significantly different from control values (p < 0.05).

ues recorded after 15 min cardiac arrest and 2 h later corresponds to a temperature drop of 5”, which is rather an underestimation of the actual decline in body temperature.

The ADC changes observed during cardiac arrest are obvioudy related to the metabolic and not to the hemodynamic disturbance, because the decline does not occur instantaneously. 5*24 In fact, using single voxel ADC measurements with 10-s time resolution Davis et al.” reported that ADC begins to decline at 2-4 min after cardiac arrest. However, although many authors suggested before that ADC changes are closely correlated with energy metabolism,4,5,26 direct comparisons of ADC maps with regional alterations of energy state have only been carried out in our laboratory.7327*28 Using bioluminescence and fluoroscopic imaging of

ATP and tissue pH, respectively, evidence was provided that normalization of ADC after 1 h global ischemia is associated with recovery of energy and acid-base homeostasis, whereas absence of postischemic ADC recovery correlates with persisting acidosis and energy depletion.’ This observation is fully in line with the present study, in which a similar relationship has been observed. However, in focal ischemia induced in rats by middle cerebral artery occlusion, the correlation between ADC and energy state was more complex.27,28 During the initial 2 h of vascular occlusion, areas with reduced ADC were consistently larger than those of ATP depletion, 28 but after 6 h occlusion or longer, the two areas merged. 27 This suggests that ADC reduction is possible in the absence of ATP depletion, although this dissociation does not persist for extended periods.

ADC recovery after cardiac arrest 0 M. FISCHER ET AL. 181

Fig. 2. Quantitative ADC maps of the brain of a cat submitted to 15-min cardiac arrest followed by successful cardiac resuscitation. Seven slices of 3 mm thickness were recorded simultaneously before and during 15-min cardiac arrest, as well as 20 min, 60 min, and 180 min after resuscitation by closed-chest cardiac massage. Note normalization of ADC at 60 min after beginning of resuscitation and the transient overshoot of ADC in brainstem and basal ganglia after 20 min.

Fig. 3. Quantitative ADC maps of the brain of a cat submitted to 15-min cardiac arrest without successful heart resuscitation. Six slices of 3 mm thickness were recorded at 20, 60, and 120 min. Note absence of ADC normalization after cardiac arrest and its further decline after unsuccessful resuscitation.

The observations of ADC changes during and after ischemia are in line with the notion that ADC reflects mainly alterations in extracellular volume,6s7 which in turn, depends on the osmolality of the tissue.29 In focal ischemia, intracellular osmolality begins to increase when blood flow decreases below the threshold of an- aerobic glycolysis, and it further rises when anoxic cell membrane depolarization results in equilibration of transmembrane ion gradients. Following cardiac arrest, both events occur in fast succession. ADC changes during complete global ischemia correlate, therefore, closer with alterations of energy state than during focal ischemia.

Our observations on the normalization of ADC and the regional energy state confirm our previous conclusion that 15min ventricular fibrillation in the anesthetized normothermic cat is compatible with homogeneous restoration of brain metabolism.’ By using the model of remotely controlled cardiac arrest and resuscitation presented here, it has become possible for the first time to follow this process with high temporal and spatial resolution. The data demonstrate that during the early resuscitation period substantial heteroge- neities exist, which, however, are reversed during the initial few hours of recirculation. This suggests that after 15min cardiac arrest followed by pneumatic vest resuscitation the previously reported recirculation disturbances l5 are reversed and do not result in persisting ADC or bioluminescence detectable metabolic alterations. However, it cannot be excluded that alterations persist at the microcirculatory level and that these may cause persisting disturbances of neurological function. In fact, Todd et a1.12,13 reported severe neurological deficits at 24 h after resuscitation from 1Zmin ventricular fibrillation, and Rosenthal et a1.r4 even after only lo-min cardiac arrest. Obviously, it has to be clarified whether these observations were made under less favor- able resuscitation conditions than in our investigation, or if the normalization of ADC and energy metabolism does not predict recovery of neurological function, as suggested by Martin et a1.30

In conclusion, the present study revealed that breakdown and recovery of cerebral energy metabolism following cardiac arrest and resuscitation can be re- liably monitored by ADC mapping. The method of the present study to induce cardiac arrest and to per- form closed-chest cardiac massage, defibrillation, and reperfusion without removing the animal from the magnet, enables the continuous monitoring of postischemic brain recovery with high temporal and spatial resolution. This provides the unique opportunity to noninvasively investigate the effects of new therapeutic strategies during the critical phase of postcardiac arrest resuscitation.

FORE

BRAI

N CO

RTEX

lwo-

. SU

CCES

SFUL

. UN

SUCC

ESSF

UL

. .

. .

. .

. .

. .

. .

20

40

80

80

100

120

140

180

18Om

in

co

CPR

RECI

RCUL

ATIO

N

BASA

L G

ANG

LIA

BRAI

NST

EM

I tr

. SU

CCES

SFUL

A UN

SUCC

ESSF

UL

r 1

1 20

40

SO

80

10

0 12

0 14

0 18

0 18

0min

co

CPR

RECI

RCUL

ATIO

N

. SU

CCES

SFUL

. UN

SUCC

ESSF

UL

co

CPR

RECI

RCUL

ATIO

N 20

40

80

80

10

0 12

0 14

0 IS

0 18

Omin

4 4

. SU

CCES

SFUL

. UN

SUCC

ESSF

UL

I I,,

, I

I I

I I

I ,

20

40

SO

80

100

120

140

160

180m

in co

CP

R RE

CIRC

ULAT

ION

Fig

. 4.

Mea

n re

gion

al (c

orte

x, b

asal

gang

lia, b

rain

stem

) AD

C v

alue

s (+

SD

) of s

even

cats

with

suc

cess

ful ca

rdia

c res

usci

tatio

n (se

e Fig

. 2),

and

of th

e an

imal

with

uns

uc-

cess

ful re

susc

itatio

n (see

Fig

. 3).

Not

e tr

ansi

ent o

vers

hoot

of A

DC

in

basa

l gan

glia

and

brai

nste

m of

suc

cess

fully

resu

scita

ted a

nim

als a

nd th

e pr

ogre

ssio

n of A

DC

dec

line

afte

r un

succ

essf

ul resu

scita

tion.

CA

: ca

rdia

c arr

est;

CP

R: c

ardi

opul

mon

ary

resu

scita

tion.

AD

C m

aps c

alcu

late

d fro

m f

our-

poin

t fit

s ar

e m

arke

d (=

4).

*Sig

nific

antly

di

ffere

nt f

rom

con

trol r

ecor

ding

s (p

< 0.

05).

,“.

. ,-

,̂. -

,,-

,” -

- *

- ,..

, - -

““.,-

-

,,,, -

- ,,.

, _ ,,

,. .,

,, .̂

,_ “.

. -

” -

,.“-

- .”

- -

,..-

- .,̂

- -

,. -

.- ,.”

-

ADC recovery after cardiac arrest 0 M. FISCHER ET AL. 789

Fig. 5. Comparison of ADC maps with the regional distribution of ATP, glucose, and lactate at 3 h after 15min cardiac arrest. In the upper row, the animal without successful resuscitation (see Fig. 3) shows low ADC and severe metabolic disturbances (depletion of ATP and glucose, high lactate), whereas in animals with successful resuscitation ADC and metabolism exhibit homogeneous recovery (high ADC, ATP, and glucose, low lactate).

Acknowledgmenf-We thank Mrs. M. Jagodnik, Mrs. U. Uh- lenkiiken and Mr. B. Rademacher for excellent technical assistance. We also gratefully acknowledge the help of Mrs. U. Beckmann with the bioluminescence technique, the art work of Mrs. I. Mtihlhaver and Mr. B. Huth, and the secretarial assistance of Mrs. D. Schewetzky and Mrs. M. Hahmann. The study was supported by the Deutsche Forschungsgemeinschaft (SFB 194/Bl).

REFERENCES

1. Siesjo, B.K. Brain Energy Metabolism. Chichester: Wiley Interscience Publ.; 1978.

2. Harreveld, A.v.; Ochs, S. Cerebral impedance changes after circulatory arrest. Am. J. Physiol. 187:180-192; 1956.

3. Hossmann, K.-A. Cortical steady potential, impedance and excitability changes during and after total ischemia of cat brain. Exp. Neurol. 32:163-175; 1971.

4. Moseley, M.E.; Cohen, Y.; Mintorovitch, J.; Chileuitt, L; Shimizu, H.; Kucharczyk, J.; Wendland, M.F.; Wein- stein, P.R. Early detection of regional cerebral ischemia in cats: Comparison of diffusion- and TZweighted MRI and spectroscopy. Magn. Reson. Med. 14:330-346; 1990.

5. Bizzi, A.; Righini, A.; Turner, R.; LeBihan, D.; DesPres, D.; Di Chiro, DG.; Alger, J.R. MR of diffusion slow- ing in global cerebral ischemia. Am. J. Neuroradiology 14:1347-1354: 1993.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

Benveniste, H.; Hedlund, L.W.; Johnson, G.A. Mech- anism of detection of acute cerebral ischemia in rats by diffusion-weighted magnetic resonance microscopy. Stroke 23~746-754; 1992. Hossmann, K.-A.; Fischer, M.; Bockhorst, K.; Hoehn- Berlage, M. NMR imaging of the apparent diffusion coefficient (ADC) for the evaluation of metabolic sup- pression and recovery after prolonged cerebral ischemia. J. Cereb. Blood Flow Metab. 14:723-731; 1994. Seo, K.; Ishimaru, S.; Hossmann, K.-A. Two-stage resuscitation of the cat brain after prolonged cardiac arrest. lntensive Care Med. 17:410-418; 1991. Hossmann, V.; Hossmann, K.-A. Return of neuronal function after prolonged cardiac arrest. Brain Res. 60: 423-438; 1973. Iijima,T.; Bauer, R.; Hossmann, K.-A. Brain resuscitation by extracorporeal circulation after prolonged cardiac arrest in cats. Intensive Care Med. 19:82-88; 1993. Safar, P.; Stezoski, W.; Nemoto, E.M. Amelioration of brain damage after 12 minutes cardiac arrest in dogs. Arch. Neurol. 33:91-95; 1976. Todd, M.M.; Dunlop, B. J.; Shapiro, H.M.; Chadwick, H.C.; Powel, H.C. Ventricular fibrillation in the cat: A model for global ischemia. Stroke 12:808-815; 1981. Todd, M.M.; Chadwick, H.S.; Shapiro, H.M.; Dunlop, B.J.; Marshall, L.F.; Dueck, R. The neurologic effects of thiopental therapy following experimental cardiac arrest. -4nesthesiology 57:76-86; 1982. Rosenthal, R.E.; Williams, R.; Bogaert, Y.E.; Getson, P.R.; Fiskum, G. Prevention of postischemic canine neurological injury through potentiation of brain energy metabolism by acetyl-L-carnitine. Stroke 23:1312-1317; 1992. Fischer, M.S.; Hossmann, K.-A. No-reflow after cardiac arrest. Intensive Care Med. 21:132-141 (1995). Sterz, F.; Safar, P.; Johnson, D.W.; Oku, K.; Tisherman, S.A. Effects of U74006F on multifocal cerebral blood flow and metabolism after cardiac arrest in dogs. Stroke 22:889-895; 1991. Halperin, H.R.; Guerci, A.D.; Chandra, N.; Herskowitz, A.; Tsitlik, J.E.; Niskanen, R.A.; Wurmb, E.; Weisfeldt, M.L. Vest inflation without simultaneous ventilation during cardiac arrest in dogs: Improved survival from prolonged cardiopulmonary resuscitation. Circulation 74: 1407-1415; 1986. American Heart Association. Guidelines for cardiopulmonary resuscitation and emergency cardiac care. Emer- gency Cardiac Care Committee and Subcommittees, American Heart Association. Part III. Adult advanced cardiac life support. JAMA 268:2199-2241; 1992. LeBihan, D.; Breton, E.; Lallemand, D.; Grenier, P.; Cabanis, E.; Laval-Jeantet, M. MR imaging of the in- travoxel incoherent motions: Application to diffusion and perfusion in neurologic disorders. Radiology 161: 401-407; 1986. PontCn, U.; Ratcheson, R.A.; Salford, L.G.; Siesjo, B.K. Optimal freezing conditions for cerebral metabo- lites in rats. J. Neurochem. 21:1127-l 138; 1973. Paschen, W.; Niebuhr, I; Hossmann, K.-A. A biolu-


nescence method for the demonstration of regional glucose distribution in brain slices. J. Neurochem. 36: 513-517; 1981

22. Paschen, W. Regional quantitative determination of Iac- tate in brain sections. A bioluminescent approach. J. Cereb. Blood Flow Metab. 5:609-612; 1985

23. Hoehn-Berlage, M.; Back, T.; Norris, D.G. Early changes in apparent diffusion coefficient of rat brain foi- lowing total circulatory arrest. MAGMA 2:39-42; 1994.

24. Davis, D.; Ulatowski, J.; Eleff, S. ; Izuta, M.; Mori, S. ; Shungu, D.; van Zijl, P.C.M. Rapid monitoring of changes in water diffusion coefficient during reversible ischemia in cat and rat brain. MRM 3 1:454-460; 1994.

25. Hasegawa, Y.; Latour, L.; Sotak, C.H.; Dardzinsky, B.J.; Fisher, M. Temperature dependent change of apparent diffusion coefficient of water in normal and isch- emit brains of rats. J. Cereb. Blood Flow Metab. 14: 383-390; 1994.

26. Mintorovitch, J.; Moseley, M.E.; Chileuitt, L.; Shimizu, H.; Cohen, Y.; Weinstein, P.R. Comparison of diffusion-

and TZweighted MRI for the early detection of cerebral ischemia and reperfusion in rats. Magn. Reson. Med. 18: 39-50; 1991.

27. Back, T.; Hoehn-Berlage, M.; Kohno, K.; Hossmann, K.-A. Diffusion nuclear magnetic resonance imaging in experimental stroke. Correlation with cerebral metabo- lites. Stroke 25:494-500; 1994.

28. Kohno, K.; Hoehn-Berlage, M.; Mies, G.; Back, T.; Hossmann, K.-A. Relationship between diffusion- weighted magnetic resonance images, cerebral blood flow and energy state in experimental brain infarction. Magn. Reson. Imaging 13:73-80; 1995.

29. Hossmann, K.-A.; Takagi, S. Osmolality of brain in cerebral ischemia. Exp. Neurol. 51:124-131; 1976.

30. Martin, G.B.; Nowak, R.M.; Paradis, N.; Rosenberg, J.; Walton, D.; Smith, M.; Eisiminger, R.; Welch, K.M. Characterization of cerebral energetics and brain pH by 31P spectroscopy after graded canine cardiac arrest and bypass reperfusion. J. Cereb. Blood Flow Metab. 10:221- 226; 1990.

imaging of the apparent diffusion coefficient for the evaluation of cerebral metabolic recovery...

Documents