ORIGINAL ARTICLES

Cerebral Blood Flow and Metabolism in Hypothermic Circulatory Arrest Craig K. Mezrow, MS, Ali M. Sadeghi, MD, Alejandro Gandsas, MD, Howard H. Shiang, DVM, Dale Levy, MD, Robert Green, MD, Ian R. Holzman, MD, and Randall B. Griepp, MD Mount Sinai Medical Center, New York, New York

Although hypothermic circulatory arrest has been accepted for use in cardiovascular operations, the potential for cerebral injury exists. The mechanism of the cerebral injury remains unclear. To address these questions we studied cerebral blood flow and metabolism. Sixteen puppies were randomly assigned to undergo either 45 or 90 minutes of hypothermic circulatory arrest after perfusiodsurface cooling to 13°C. Cerebral blood flow, cerebral oxygen and glucose metabolism, and cerebral vascular resistance measurements were obtained at 37T, 13"C, 10 minutes after reperfusion, 30°C and 2 and 4 hours after hypothermic circulatory arrest. No neurologic or behavioral changes were observed in any of the

long-term survivors (11/16). Metabolic and cerebral blood flow data did not differ between groups. Cerebral blood flow was significantly lower in the late postarrest measurements, whereas oxygen and glucose consumption had returned to baseline values. In the presence of low cerebral blood flow and high cerebral vascular resistance it is notable that control levels of oxygen consumption were attained by abnormally high oxygen extraction. These data strongly suggest a vulnerable interval after hypothermic circulatory arrest in which cerebral metabolism is limited by cerebral blood flow.

H

increased extraction of oxygen and glucose. We believe that this constitutes a period of increased vulnerability to cerebral injury after HCA.

ypothermic circulatory arrest (HCA) has facilitated the surgical repair of a wide variety of cardiovascular anomalies. The safety of HCA is largely due to a temperature-dependent reduction of metabolic rate. The brain, the organ most sensitive to ischemic injury, is the limiting factor with regard to the duration of HCA. Although clinical experience suggests that intervals of HCA of 60 minutes or shorter do not result in overt cerebral injury, the results of a number of investigations continue to raise questions about the safety of HCA duration exceeding 45 minutes. Because many of the operations requiring use of HCA in infancy frequently exceed 45 and even 60 minutes, a reexamination of the cerebral consequences of HCA seems warranted. The present report is part of a larger study in which we are seeking to correlate intraoperative hemodynamic and metabolic variables with clinical neurological outcome after HCA. The study is being carried out in inbred weanling puppies to minimize the impact of genetic variability and to simulate as closely as possible the clinical circumstances involved in cardiac operations using HCA in infants. Our results after intervals of HCA of 45 and 90 minutes demonstrate a decrease in cerebral blood flow (CBF) and increase in cerebral vascular resistance (CVR) occurring several hours after HCA during which baseline metabolic rates are maintained only by Presented at the Twenty-eighth Annual Meeting of The Society of Thoracic Surgeons, Orlando, FL, Feb 3-5, 1992. Address reprint requests to Dr Mezrow, Department of Cardiothoracic Surgery, Mount Sinai Medical Center, PO Box 1028, New York, NY 10029.

0 1992 by The Society of Thoracic Surgeons

(Ann Thorac Surg 1992;54:609-16)

Material and Methods Sixteen weanling beagles (Marshall Farms, North Rose, NY), 3 to 4 months of age, weighing 6 to 7 kg, were randomly assigned to one of two experimental groups, either to undergo 45 or 90 minutes of HCA. Each experimental group underwent preoperative neurologic and behavioral assessment, invasive intraoperative measurements, and a postoperative neurologic and behavioral evaluation.

Anesthesia Animals were pretreated with glycopyrrolate(0.025 mgkg), anesthetized with fentanyl(Z5 to 50 pg * kg-' h-') and isoflurane (0.40; carbon dioxide tension, 35 to 40 mm Hg) and paralyzed with pancuronium. This regimen, in which blood gases were monitored according to so-called alphastat principles, is similar to that used clinically in infants, and was selected to maintain conditions that would seek to minimize changes in CBF and oxygen and glucose metabolism. Arterial oxygen tension was maintained greater than 100 mm Hg. Lead I1 was used for electrocardiographic monitoring. Temperature probes were placed in the esophagus and rectum, and an indwelling catheter was inserted in the bladder. Femoral artery and vein cannulations were performed for monitoring purposes.

-

0003-4975/92/$5.00

610

MEZROW ET AL CBF AND METABOLISM IN HCA

Pre-Bypass

I-

1992;54:609-16

Bypass Coolin0

I

I

Ann Thorac Surg

Circulatory Arrest I I Periods :(45 or SO min)l

:

I

I

IBas\e'lne I

I

Bypass Warming

I

I

: I

2 hrs . -after . -

y-

I

I I

Post-Bypass I

I I

I

37 after 10 min H C AIh

27

HCA

4 hrs after

clcn .-. . ,

G

1

7

I I

I I

r:7

I

I I

13

I

I

- 2

- 1

I I

I

I

I

-3

-4

-6

-5

-7

Approximate time (hr4

Fig 1. Measurement time points. (HCA

=

hypothermic circulatory arrest.)

Sagittal Sinus Cannulation Sagittal sinus cannulation was performed before cannulation for cardiopulmonary bypass (CPB). A midline scalp incision was made and the underlying periosteum removed to facilitate identification of the coronal and sagittal sutures. Under 2.5X magnification, a 3-mm cutting burr was used to remove the bone over the sinus. A 24-gauge catheter was inserted into the sagittal sinus to permit both sampling of cerebral venous blood and monitoring of cerebral venous pressure.

Cardiopulmonary Bypass, Cooling, and Rewarming After heparinization (300 IU/kg), nonpulsatile CPB was instituted employing single-cannula drainage of the right atrium with return of the arterial perfusate through a cannula positioned in the ascending aorta. A cannula was passed from the left atrium into the left ventricle to permit decompression of the left heart during CPB. Surface cooling was achieved using both a cooling blanket and ice packs around the head. Membrane oxygenators (VPCML plus; Cobe Laboratories, Inc, Lakewood, CO) were primed with a hemodilute solution containing homologous blood (universal donor), 5% albumin solution, furosemide (1mg/kg), heparin (2,000 IU), 1%dextrose in 0.9% saline solution, and KCl (1 mEq/kg). Cardiopulmonary bypass was established at 75 to 100 mL * kg-' * min-', and hematocrit was maintained between 0.22 and 0.28. Employing the principles of alpha-stat management, pH

Table 1. Comparison of Perfusion Cooling and Rewarming Time" Group 45 min HCA (n 90 min HCA (n a

= =

8) 8)

Weight

(kg)

CPB Cooling (min)

CPB Warming (min)

6.8 2 0.5 7.0 0.4

53 2 2 56 2

52 5 3 53 2 2

*

*

Values are shown as mean 2 standard error.

CPB = cardiopulmonary bypass; arrest.

HCA

=

hypothermic circulatory

during cooling was maintained at 7.4 & 0.05 and carbon dioxide tension at 35 to 40 mm Hg, uncorrected for temperature. Cooling was carried out to an esophageal temperature of 13°C. The total period of perfusion cooling varied from 50 to 60 minutes. Circulatory arrest was then established and the perfusate drained into the oxygenator reservoir. After the selected interval of HCA (45 or 90 minutes), perfusion/surface rewarming was carried out (CPB at 75 to 100 mL * kg-' min-' for 50 to 60 minutes) to an esophageal temperature of more than 36°C and the animals then weaned from CPB. When necessary, cardiac defibrillation was performed after administration of lidocaine (1 mg/kg). After decannulation, protamine sulfate (5 mg/kg) was administered to reverse heparinization.

-

Cerebral Blood Flow and Cerebrovascular Resistance Cerebral blood flow was measured using radionuclidelabeled microspheres as originally described by Rudolph and Heymann [l].Approximately 0.5 to 1.5 x lo6 microspheres 15 & 0.5 pm in diameter labeled with Ru1O3, Sn113, Cr5', C O ~Nb95, ~ , and Sc46(New England Nuclear, Wilmington, DE) were injected and flushed with 5 mL of saline solution into a transthoracic left ventricular catheter before and after CPB, and into the arterial cannula during CPB. Blood reference samples were withdrawn from the femoral arterial line at a constant rate (2.29 mL/min) with a Harvard withdrawal pump beginning 15 seconds before microsphere injection and ending 105 seconds after injection. After postoperative neurologic and behavioral evaluation (details described later), the animals were anesthetized and sacrificed using sodium pentobarbital (30 mg/kg) and KC1 (6 mEq/kg). In all animals (including those who succumbed), each brain was removed and weighed, and radionuclide determination was made using a gamma counter (Auto-Gamma, Packard Instruments). Analyses were carried out by computer solution of multiple simultaneous linear equations (Compusphere; Packard Instruments, Downer's Grove, IL). Cerebral blood flow was calculated using the following equation:

611

MEZROW ET AL CBF AND METABOLISM IN HCA

Ann Thorac Surg 1992;54:609-16

Table 2. Acid-Base, Blood Gas, and Hemodynamic Variables" Baseline (37°C)

Variable ______~

~

PHa 45 min HCA 90 min HCA PaO, (mm Hg) 45 min HCA 90 min HCA PaCO, (mm Hg) 45 rnin HCA 90 min HCA Hematocrit 45 min HCA 90 min HCA Values are shown as mean

HCA

=

X

30°C

2 h After HCA

4 h After HCA

*

35.5 35.3

0.4 0.5

37.2 t 0.9 36.8 t 0.1

36.5 C 0.3 36.4 t 0.3

12.8 t 0.1 12.6 2 0.1

16.2 t 0.5 17.1 t 0.5

30.7 t 0.6 29.8 t 0.2

88 2 8 87 2 5

52 2 5 53 t 4

55 t 4 63 t 4

57 2 4 46 t 6

81 2 5 83 t 2

93 f 5 97 t 2

7.35 t 0.02 7.33 2 0.01

7.37 f 0.01 7.37 +. 0.02

7.38 t 0.02 7.37 t 0.02

7.35 t 0.01 7.31 t 0.14

7.37 t 0.02 7.38 t 0.11

7.39 f 0.02 7.37 t 0.01

594 t 20 615 t 12

453 t 61 437 t 26

486 t 48 623 t 54

579 t 62 638 t 16

633 k 16 634 C 13

38 t 1 39 t 1

35 2 2 34 t 1

36 +. 1 32 t 2

33 C 1 32 2 2

36 % 2 35 t 1

0.36 t 0.01 0.36 t 0.01

0.27 t 0.01 0.25 .t 0.01

0.23 t 0.01 0.21 t 0.01

0.24 f 0.01 0.23 t 0.01

0.33 t 0.02 0.30 2 0.02

?

286 407

f

55

* 31

2 f

37 39

& f

1 1

0.35 C 0.02 0.33 t 0.02

standard error. MAP

hypothermic circulatory arrest;

CBF (mL 100 g -

10 min After HCA

~

Esophageal Temperature ("C) 45 min HCA 90 min HCA MAP (mm Hg) 45 min HCA 90 min HCA

a

13°C

min - I)

=

=

mean arterial pressure;

PaCO,

=

carbon dioxide tension;

PaO, = oxygen tension

bolic rate of oxygen (CMRO,) and cerebral metabolic rate of glucose (CMRGlu) were determined as follows:

(cerebral tissue counts

rate of withdrawal) x 100/(counts in reference sample

CMR02 (mL * 100 g -

x brain weight).

x (arterial 0

Cerebral vascular resistance was calculated by using the equation:

2

- rnin

- I)

= CBF

content - sagittal sinus 0

CMRGlu (mg * 100 g -

min - I)

2

content)/100.

=

CBF

x (arterial glucose - sagittal sinus glucose)/100.

CVR(mmHg~mL-'-100g-'~min-')=(MAP

Cerebral Metabolism

Arterial and venous blood pH, oxygen tension, carbon dioxide tension, hematocrit, and oxygen content were measured using a Ciba-Corning Diagnostics Corporation analyzer (model 288; Medfield, MA), and glucose levels were determined using a YSI Inc analyzer (model 2300; Yellow Springs, OH).

Sagittal sinus and arterial samples were obtained simultaneously for calculation of both cerebral oxygen extraction (arteriovenous oxygen content difference) and glucose extraction (arteriovenous glucose difference). Cerebral meta-

Measurements of CBF, CVR, CMRO,, and CMRGlu were made at six time points during the experiments (Fig 1):(1)

- MSSP)/CBF,

where MAP = mean arterial pressure and MSSP = mean sagittal sinus pressure.

Study Protocol

Table 3 . Cerebral Blood Flow and Vascular Resistance" Perfusion Rewarming Baseline (37°C)

Variable CBF (mL * 100 g-' . min-') 45 min HCA 90 rnin HCA CVR (mm Hg . mL-' * 100 g-' 45 min HCA 90 min HCA

=

cerebral blood flow;

10 min After HCA

90 18 74 t 9

41 t 9b 36 f 7b

65 f 11 87 2 30

1.0 f 0.1 1.2 2 0.2

1.3 0.1 1.6 2 0.2

*

. min-')

* Values are shown as mean +- standard error.

CBF

13°C

*

0.8 2 0.1 0.9 t 0.1

30°C 93 t 10 106 t 18

*

0.6 O . l b 0.5 t O.lb

p < 0.05 versus baseline (37°C).

CVR = cerebral vascular resistance;

HCA

=

hypothermic circulatory arrest.

2 h After HCA

4 h After HCA

31 41

f 9'

40 f lob 30 C 2'

2.6 2 0.3b 2.0 f 0.4b

2.7 t 0.3b 3.1 2 0.3'

f

3b

612

MEZROW ET AL CBF A N D METABOLISM IN HCA

Ann Thorac Surg 1992;54:609-16

Fig 2. Cerebral blood flow (CBF) and vascular resistance (CVR) (mean values C standard error). (HCA = hypothermic circulatory arrest; *p < 0.05 versus baseline by analysis of variance.)

13°C

baseline: at 37°C before CPB; (2) after cooling, immediately before HCA at 13°C; (3) 10 minutes after the end of HCA; (4) during rewarming at 30°C (approximately 20 minutes after the end of HCA); (5) 2 hours after the end of HCA (after termination of CPB, at 37°C with closed thorax); and (6) 4 hours after the end of HCA.

Behavioral and Neurological Assessment Puppies were examined preoperatively and observed for 1 week postoperatively for evaluation of behavioral and neurological sequelae of HCA. They were evaluated daily by a veterinarian unaware of the experimental design to determine whether behavioral changes were present (mental status, appetite, affect). Each animal also underwent neurological examination including evaluation of gait, reflexes, and cranial nerve responses.

Statistical Analysis All results are expressed as the mean +- standard error. A p value less than 0.05 as determined by analysis of variance was accepted as statistically significant.

Perioperative Manag emen t All animals have received humane care in compliance with the ”Guide for the Care and Use of Laboratory Animals” published by the National Institutes of Health (NIH publication No. 85-23, revised 1985). The protocol for these experiments was approved by the Mount Sinai Institutional Animal Care Committee, and the animals were housed in the institution’s NIH-approved animal facility preoperatively and postoperatively.

Results Demographic data are displayed in Table 1. There was no difference between the groups of puppies undergoing 45 or 90 minutes of HCA with respect to weight or duration of CPB. Table 2 shows the controllable variables measured for each animal at the six times that data were collected for calculation of CBF and cerebral metabolism in each group.

10 rnin Post HCA

30%

2 hr Post HCA

4 hr Post HCA

As intended by the design of the study, there were no significant differences between animals in the two groups in terms of temperature, mean arterial pressure, arterial pH, blood gases, or hematocrit at any of the time points.

Long-Term SurvivallNeurological Evaluation Although all animals included in the study survived long enough to complete the intraoperative protocol and were successfully weaned from CPB, 7 of 8 animals survived for 1 week after 45 minutes of HCA, but only 4 of 8 survived after 90 minutes of HCA. All deaths were attributed to noncardiogenic pulmonary congestion. No neurologic or behavioral changes were present in any of the surviving animals.

Cerebral Blood Flow and Cerebral Vascular Resistance The values for CBF and CVR for the six time points for animals undergoing 45 minutes and those undergoing 90 minutes of HCA are shown in Table 3. There being no differences between the two groups, the data for both groups were combined in Figures 2 and 3A. There was a significant decrease in CBF at 13°C before HCA compared with baseline values before CPB at 37°C. During rewarming, CBF very rapidly returned to baseline values, but then fell markedly by 2 hours after the end of HCA and remained significantly lower than baseline values 4 hours after HCA (see Fig 2). Cerebral vascular resistance did not show a significant increase upon cooling but decreased during rewarming: a difference from baseline was significant at the 30°C timepoint (Fig 3B; see Table 3). By 2 hours after HCA, however, CVR was markedly increased, and it remained significantly higher than baseline 4 hours after HCA.

Cerebral Oxygen Metabolism There were no significant differences between groups with different durations of HCA with regard to sagittal sinus oxygen saturation, oxygen extraction, or CMRO, (Table 4). Hence, the values for both groups were combined in Figures 3C, 3D, and 4. As might have been predicted, oxygen extraction and

T

100

..

613

MEZROW ET AL CBF A N D METABOLISM IN HCA

Ann Thorac Surg 1992;54:609-16

c 80E

w

0 0

E

.

60-

40 ~

201 Baeellne

13°C

10 rnln Post HCA

30%

2 hr

Post HCA

~~

Easellne

13°C

8.SdlM

13°C

10 rnln Post HCA

30°C

2 hr

10 rnln

30°C

2 hr

Post HCA

4 hr

4 hr

D

A

Post HCA

B

Post HCA

Baseline

13°C

10 mln P o d HCA

30°C

2 hr

4 hr Post HCA

4 hr

Post HCA

E

.I 0

Pod HCA

T

I

C

F Fig 3. (A) Cerebral blood flow. ( B ) Cerebral vascular resistance. (Q Oxygen extraction. (D)Cerebral metabolic rate of oxygen. ( E ) Glucose extraction. (F) Cerebral metabolic rate of glucose. All values are shown as mean standard error. (HCA = hypothermic circulatory arrest.)

*

consumption were markedly reduced by cooling and remained lower than baseline values as rewarming began. By 2 hours after HCA, however, at a time when oxygen consumption had returned to levels not significantly different from baseline values, oxygen extraction had risen to levels significantly higher than control values, correlating with significantly depressed sagittal sinus oxygen saturations.

cooling and early rewarming (see Fig 3E). As was the case with oxygen metabolism, by 2 hours after HCA CMRGlu had returned to levels not significantly different from baseline values but glucose extraction had risen to levels significantly higher than baseline values; the levels of glucose extraction remained significantly elevated 4 hours after HCA (see Figs 3E, 3F).

Cerebral Glucose Metabolism

The relationship between the changes in CBF, CVR, and cerebral metabolism can perhaps best be summarized by taking the ratio CBF/CMRO, (Fig 5). If the baseline value of this ratio is assumed to represent normal cerebral autoregulation, ie, an optimal relationship between CBF

No notable differences between the experimental groups were present with regard to glucose extraction and CMRGlu (Table 5). As was seen with oxygen consumption, glucose extraction was markedly reduced during

Summary of Results

614

MEZROW ET AL CBF AND METABOLISM IN HCA

Ann Thorac Surg 1992;54:609-16

Table 4. Cerebral Oxvgen Saturation and Metabolism“ Baseline (370C)

SS 0, sat (%) 45 min HCA 90 min HCA 0, extraction (mL/dL) 45 min HCA 90 min HCA CMRO, (mL . 100 g-’ . min-I) 45 min HCA 90 min HCA a

Values are shown as mean

13°C

29 t 3’ 33 t 2’

24 t 1’ 25 t 2b

6.9 t 0.8 6.4 t 0.6

1.5 t 0.3’ 1.3 f 0.3’

1.6 t 0.2’ 1.7 t 0.2b

2.9 ? 0.3’ 3.1 t 0.3b

12.4 t 0.5’ 11.1 t 0.5’

14.1 t 0.7 13.4 2 0.8

5.6 t 1.2 4.5 t 0.6

0.5 t 0.1’ 0.4 t O . l b

1.0 f 0.1’ 1.5 2 0.8’

2.7 t 0.4 3.1 t 0.5

3.8 t 0.3 4.6 t 1.3

5.8 t 2.1 4.0 t 0.4

p < 0.05 versus baseline (37°C). HCA

=

hypothermic circulatory arrest;

Comment Although there is some uncertainty concerning the safe duration of HCA, there is almost universal agreement that when the duration of HCA exceeds 60 minutes there is increased risk of cerebral injury. It is not known whether the cerebral injury that can occur as a consequence of HCA takes place during the period of circulatory arrest or during reperfusion. The mechanism of injury also remains unknown. To address these questions we investigated the impact of ”safe” (45 minutes) and potentially ”unsafe” (90 minutes) intervals of HCA on CBF, CVR, CMRO,, and CMRGlu. We also sought to correlate these intraoperative variables with neurological outcome. We were surprised by the fact that there was no evidence of adverse neurological outcome in any of the

loot

2 hr

0, extraction

=

arteriovenous oxygen content difference;

animals surviving 90 minutes of HCA. The animals who died did not live long enough to ascertain whether they had any overt neurological deficits. Moreover, there were no significant differences in measured variables between those puppies who survived and those who died after operation in both HCA groups. All deaths occurred within 24 hours of CPB and were attributed to noncardiogenic pulmonary congestion, a complication of our CPB technique. Given the absence of any significant differences between the two groups either in neurological or behavioral outcome or in intraoperative measures of cerebral blood flow or metabolism, we suggest that no obvious cerebral injury is discernible under conditions of 90 minutes of HCA at 13°C employing combined surface and perfusion cooling and rewarming. We recognize that clinical studies that have shown effects of longer intervals of HCA have generally shown subtle impairment of cognitive function, and that without the ability to measure intelligence in our animal model, we would need sophisticated quantitative electroencephalographic analysis and careful postmortem histological evaluation of the brain before firmly concluding that no cerebral damage has occurred even after the longer intervals of HCA used in this study. Although histopathological assessment of these animals is not complete, past experience with this model suggests that when no clinical evidence of cerebral injury is present unequivocal histopathological changes are unlikely to be found.

*

30°C

78 t 3 79 2 3

HCA

94 f 2’ 97 t lb

* standard error.

10 rnl” Post HCA

4 h After

95 t 2’ 96 t I’

and cerebral metabolism, then CBF is excessive-this has been dubbed “luxury perfusion” during cooling and immediately upon rewarming. Late after HCA, the ratio CBFKMRO, falls well below its optimal value so that normal or slightly increased rates of cerebral metabolism can only be maintained by means of increased extraction of oxygen and glucose. The inappropriately high CBF during cooling is reflected by-or perhaps a consequence of-the unchanged CVR during cooling, and the decrease in CBF at the late time points is correlated with a marked increase in CVR.

*

2 h After

HCA

30°C

67 t 5 73 t 3

CMRO, = cerebral metabolic rate of oxygen; SS 0, sat = sagittal sinus oxygen saturation.

inn 1

10 min After HCA

*

T

4 hr

Post HCA

Fig 4 . Cerebral venous oxygen saturation (mean values f standard error). (HCA = hypothermic circulatory arrest; *p < 0.05 versus baseline by analysis of variance.)

Fig 5. Relationship of cerebral blood flow (CBF) to cerebral metabolic rate of oxygen (CMROZ) (mean values t standard error). (HCA = hypothermic circulatory arrest; “p < 0.05 versus baseline by analysis of variance.)

1

MEZROW ET AL CBF AND METABOLlSM IN HCA

Ann Thorac Surg 1992;54:609-16

615

Table 5. Cerebral Glucose Metabolism“ Baseline (37°C)

13°C

7 f 1 8 f 1

0.3 f 0.2’ 0 t 0’

5.1 t 0.4 5.6 2 0.8

0.1 t 0.1’ 0 t Oh

Variable Glucose extraction (mg/dL) 45 min HCA 90 min HCA CMRGlu (mg . 100 g-’ . min-’) 45 min HCA 90 min HCA a

Values are shown as mean t standard error.

CMRGlu = cerebral metabolic rate of glucose;

10 min After HCA 1 f 0.4’ 1 f 1’

0.3 0.7

* 0.2’ ?

0.5’

30°C

2 h After HCA

4 h After HCA

6 f 2 10 t 3

12 f 2‘ 11 2 1

15 +. I’ 17 f 2’

3.4 t 0.5 4.9 f 1.6

6.2 f 2.0 5.0 +- 0.6

6.0 8.9

f f

2.3 2.6

p < 0.05 versus baseline (37°C). glucose extraction

=

arteriovenous glucose difference;

The decrease in CBF and cerebral metabolism during cooling is in accord with theoretical predictions and the experimental work of others [2, 31. The marked reduction in oxygen and glucose metabolism just before HCA at 13°C add to available evidence demonstrating the theoretical safety of circulatory arrest at this temperature. The earliest measurements of oxygen and glucose metabolism after the interval of HCA are also reassuringly low, suggesting that cerebral metabolism is relatively quiescent during even the longer interval of circulatory arrest. During rewarming (30°C), both oxygen and glucose metabolism returned to near baseline values but oxygen extraction remained depressed. This may reflect a persistent effect of low temperature on glucose and oxygen extraction or the fact the CBF is actually excessive relative to both temperature and metabolic requirements so that less extraction is required to meet metabolic demands. The increase in CVR and decrease in CBF after HCA are in agreement with the work of others using a variety of methods for the study of CBF [2-51. The contribution that our study makes is in showing that this decreased CBF appears to be inappropriate in terms of the metabolic demands of the brain during its recovery from HCA. The discordance between CBF and cerebral metabolism suggests that there is a disturbance in cerebral autoregulation after HCA, and by monitoring these variables longer after HCA than has been done previously, we have shown that the inappropriately depressed CBF and increased CVR last at least 4 hours. We cannot be sure from our data whether even higher metabolic rates would not be better for optimal cerebral recovery, or whether at some point beyond 4 hours after HCA the decrease in CBF becomes so severe that a further increase in oxygen extraction cannot take place to maintain the appropriate metabolic rate. In this study, the decrease in CBF observed in the late postoperative period was unrelated to any change in mean arterial pressure or other adverse postoperative event and was apparently adequately compensated for by increased extraction of both oxygen and glucose to main-

HCA

=

hypothermic circulatory arrest.

tain the cerebral metabolic rate at baseline levels. It seems likely, however, that any decrease in arterial pressure, decreased hematocrit, o r hypoxia might impair oxygen delivery during this interval of reduced CBF. We are persuaded that this late reperfusion period is a time when the brain is particularly vulnerable to injury because even under optimal circumstances it is already relying on reserve mechanisms to maintain cerebral metabolism. Our study suggests that an interval of impaired CBF occurs even after ”safe” intervals of HCA. Further studies will be needed to define the length of the vulnerable interval after HCA and to determine whether this period of increased potential for cerebral injury is also present after clinical alternatives to HCA such as low-flow moderately hypothermic perfusion. Supported by grant HL-45636-02 from the National Heart, Lung, and Blood Institute. We thank Rashin Nouranifar and Luigi Compere for their technical assistance in the conduct of these experiments.

References 1. Rudolph AM, Heymann MA. The circulation of the fetus in utero: methods for studying distribution of blood flow, cardiac output and organ blood flow. Circ Res 1967;21:16=4. 2. Greeley WJ, Kern FH, Ungerleider RM, et al. The effect of hypothermic cardiopulmonary bypass and total circulatory arrest on cerebral metabolism in neonates, infants, and children. J Thorac Cardiovasc Surg 1991;101:783-94. 3. Perna AM, Gardner TJ, Tabaddor K, Brawley RK, Gott VL. Cerebral metabolism and blood flow after circulatory arrest during deep hypothermia. Ann Surg 1973;178:95-101. 4. Settergren G, Ohqvist G, Lundberg S, Henze A, Bjork VO, Persson B. Cerebral blood flow and cerebral metabolism in children following cardiac surgery with deep hypothermia and circulatory arrest. Clinical course and follow-up of psychomotor development. Scand J Thorac Cardiovasc Surg 1982; 16:209-15. 5. Hillier SC, Burrows FA, Bissonnette B, Taylor RH. Cerebral hemodynamics in neonates and infants undergoing cardiopulmonary bypass and profound hypothermic circulatory arrest: Assessment by transcranial Doppler sonography. Anesth Analg 1991;72:723-8.

DISCUSSION DR RICHARD A. JONAS (Boston, MA): I want to congratulate you on an excellent presentation of a very interesting study. In Boston we have also developed a model of circulatory arrest in

neonatal piglets and have been looking at similar parameters. We concur with your finding that cerebral metabolic rate determined by oxygen extraction has returned to baseline within about 3

616

MEZROW ET AL CBF A N D METABOLISM IN HCA

hours after 60 minutes of circulatory arrest at a nasopharyngeal temperature of 15°C. I have two questions. You said in your conclusion that you believed there was a potentially vulnerable period where decreased pressure might result in inadequate cerebral perfusion. This is equivalent to a loss of cerebral autoregulation, in other words, pressure passive cerebral blood flow. Is that in fact what you found? You stated that you found that there was some uncoupling of cerebral metabolic rate and cerebral blood flow, but that is not the same as loss of pressure/flow autoregulation. MR MEZROW Our findings after cardiopulmonary bypass revealed that there is an interval of cerebral hypoperfusion in which cerebral metabolism is maintained by abnormally high oxygen extraction. We refer to this period as “the vulnerable period” because if there were to be any further decrease in cerebral perfusion or oxygen carrying capacity, further increases in cerebral oxygen extraction to maintain cerebral metabolism may not be possible and cerebral injury may ensue. DR JONAS: Did you specifically look at cerebral blood flow versus systemic blood pressure? MR MEZROW It was not the intent of this study to investigate the relationship of arterial blood pressure to cerebral blood flow. All measurements in this study were performed with similar mean arterial pressures and hematocrit values. DR JONAS: But these animals were off bypass MR MEZROW All postbypass measurements were performed with mean arterial pressures similar to baseline values. DR JONAS: My second question relates to pH strategy. In 1985 we changed clinically from pH-stat to alpha-stat strategy. We subsequently noted several cases of choreoathetosis after deep hypothermic circulatory arrest for complex anomalies. A recent retrospective neurodevelopmental study at Children’s Hospital has indicated that the use of alpha-stat with circulatory arrest has been associated with a worse cognitive outcome relative to the more acidotic strategy that we were using previously. Have you any information regarding the influence of pH strategy in your study? MR MEZROW We employed alpha-stat blood gas management during the conduct of this study as it is performed clinically at our institution. This study was designed to mimic the clinical practices at our institution as closely as possible. We have no data to suggest alpha-stat blood gas management to be superior to pH-stat blood gas management, but we believe there are numerous studies to suggest this. DR ROSS M. UNGERLEIDER (Durham, NC): I would like to congratulate you on this study, and especially for focusing attention on this area that is important for neonatal heart surgeons. I also congratulate you for performing behavioral studies in puppies. I think that is remarkable! I do have a couple of questions about your model. Number one, you do not really have any control animals, that is, animals that were subjected to cardiopulmonary bypass and hypothermia without total circulatory arrest and rewarmed, and it leads me to question whether the changes we are seeing are entirely due to circulatory arrest or to the model itself. Do you have those data?

Ann Thorac Surg 1992;54:609-16

Second, I am interested in your initial statement that you chose two periods of circulatory arrest, 45 minutes being so-called “safe” and 90 minutes being “unsafe,” and yet you did not see any differences between these groups. Are we supposed to assume then that there is no difference between the ”safe” versus “unsafe” times? By combining these data, have you obscured valuable information? More importantly, from our own studies, I am surprised you found no difference. Finally, I just have a question regarding the cerebral blood flow and metabolic changes. In babies we have found that there is a deficit in CMRO, immediately after hypothermic circulatory arrest. It does recover after a period of time. I wonder if the fact that you waited 2 hours until you obtained your first data point may mean that you might have missed depressed cerebral metabolism had you obtained data sooner. In fact, when I look at the decreased cerebral blood flow I wonder immediately, are your cardiac output and hemodynamics identical to the prearrest time or is blood flow diminished to the entire animal, or if cerebral blood flow was diminished, where is the increased flow going? But overall I would like to congratulate you on an interesting study. MR MEZROW. With regard to your question of controls, we did not perform a control group of animals not undergoing periods of circulatory arrest. This study was designed for every animal to be its own control with baseline measurements. We are currently investigating cerebral blood flow and metabolism after periods of hypothermic low-flow cardiopulmonary bypass. With regard to your question of the “safe” and ”unsafe” periods of hypothermic circulatory arrest investigated, we were surprised by the fact that there was no evidence of adverse neurological outcome in any of the animals surviving 90 minutes of hypothermic circulatory arrest. There were no significant differences in measured variables between those animals subjected to 45 or 90 minutes of hypothermic circulatory arrest. Without the sophisticated measurements of quantitative electroencephalography and histopathologic evaluations of the brain we cannot conclude with any certainty that there was absence of cerebral injury in animals subjected to 90 minutes of hypothermic circulatory arrest. With regard to your question of CMRO, immediately after hypothermic circulatory arrest, our measurement time points after circulatory arrest were 10 minutes after circulatory arrest, 30°C esophageal temperature, and 2 and 4 hours after circulatory arrest. Approximately 1 hour separates the 30°C and 2 hour postarrest measurements. Our 30°C data of decreased CMRO, concur with those you describe in your babies immediately after circulatory arrest. We believe this is due to the hypothermic state of the patient. With regard to your question of cardiac output and hemodynamics after circulatory arrest compared with before cardiopulmonary bypass, all measurements were made with similar hemodynamic variables such as mean arterial pressure and hematocrit values. We performed cardiac output measurements in several animals in this study and found no significant differences in prebypass and postbypass measurements, suggesting that postbypass cerebral hypoperfusion is not due to altered cardiac function.