Cardiopulmonary bypass and the blood-brain barrier

Cardiopulmonary bypass and the blood-brain barrier An experimental study The diffuse inflammation produced by cardiopulmonary bypass might disrupt the blood-brain barrier and lead to the transient neurologic dysfunction occasionaUy seen after cardiac operations. To evaluate this possibility, blood-brain barrier integrity was measured by carbon 14-aminoisobutyric acid tracer technique after 2 hours of cardiopulmonary bypass in piglets. Six animals were cooled to 28° C on cardiopulmonary bypass and then rewarmed to 38° C before carbon 14-aminosisobutyric acid was injected intraarteriaUy. A control group of six animals underwent median sternotomy and heparinization but were not placed on cardiopulmonary bypass. Blood-to-brain transfer coefficients for carbon 14-aminosisobutyric acid were calculated for multiple brain regions; higher coefficients reflect greater flux of carbon 14-aminosisobutyric acid and suggest loss of blood-brain barrier integrity. The brain regions examined and their transfer coefficients (cardiopulmonary bypass versus control mean ± standard error of the mean m1/gm/min) were middle cerebral artery territory cortex (0'()o32 ± 0.0002 versus 0.0030 ± 0.0002; p = 0.42), diencephalon (0.0031 ± 0.0003 versus 0.0029 ± 0.0002; p = 0.50), midbrain (0.0028 ± 0.0002 versus 0.0027 ± 0.0002; p = 0.86), cerebeUum (0.0036 ± 0.0003 versus 0.0029 ± 0.0002; p = 0.22), and spinal cord (0.0035 ± 0.0003 versus 0.0041 ± 0.0008; p = 0.48). There were no significant differences in transfer coefficients between animals placed on cardiopulmonary bypass and control animals in any brain region examined. The pituitary gland lacks a blood-brain barrier and had a correspondingly high coefficient in control animals and those undergoing cardiopulmonary bypass (0.077 ± 0.012 versus 0.048 ± 0.008; p = 0.07). Two hours of moderately hypothermic cardiopulmonary bypass does not disrupt the blood-brain barrier. (J THORAC CARDIOVASC SURG 1992;104:1110-5)

A. Marc Gillinov, MD,a Elizabeth A. Davis, BS,a William E. Curtis, MD,a Charles L. Schleien, MD,b Raymond C. Koehler, PhD,b Timothy J. Gardner, MD,a Richard J. Traystman, PhD,b and Duke E. Cameron, MD,a Baltimore, Md.

AthOugh mortality associated with cardiac operations has decreased in the last two decades, there continues to be a small but significant neurologic morbidity associated with use of cardiopulmonary bypass (CPB).1-4 Neurologic deficits, ranging from transient confusion to severe strokes, occur with a frequency of 0% to 60% after proFrom the Departments of Surgery, Division of Cardiac Surgery; and Anesthesiology and Critical Care Medicine," The Johns Hopkins Medical Institutions, Baltimore, Md. Supported in part by United States Public Health Service National Institutes of Health grant NS20020 and HL 19414-IOAI. Received for publication June 3, 1991. Accepted for publication Nov. 18, 1991. Address for reprints: Duke E. Cameron, MD, The Johns Hopkins Hospital, Blalock 618, 600 N. Wolfe St., Baltimore, MD 21205.

12/1/35766

I1I0

cedures involving CPB. I-7 These neurologic complications are usually ascribed to cerebral hypoperfusion, microemboli, or macroemboli.!- 5, 7-10 The relative importance of each mechanism remains unclear. I, 8,10 CPB causes complement activation and a resultant generalized inflammatory reaction that Kirklin II termed "the postperfusion syndrome." In the postperfusion syndrome there is increased vascular permeability and interstitial edema, II, 12 with a measurable increase in extravascular lung water.P A similar increase in brain capillary permeability would result in cerebral edema and might account for some neurologic deficits associated with CPB. To examine the effects of CPB on the blood-brain barrier (BBB), we used a well-established tracer technique (carbon 14-aminoisobutyric acid (AIB)I4-18 to test the hypothesis that 2 hours of hypothermic CPB disrupts the BBB.

Volume 104

Number 4 October 1992

CPB and BBB

1111

Table I. Arterial blood gases, MAP, and hemoglobin in control piglets and in piglets undergoing CPB Arterial

Pao2

Paco,

Hemoglobin

MAP

pH

(mmHg)

(mmHg)

(gmfdl)

(mmHg)

Baseline Control CPB

7.42 ± 0.05 7.40 ± 0.02

246 ± 38 224 ± 22

34.4 ± 1.6 36.2 ± 2.7

10.4 ± 0.5 10.9 ± 1.3

79.8 ± 4.9 71.7 ± 5.8

15 min Control CPB

7.44 ± 0.04 7.34 ± 0.03

288 ± 44 498 ± 101

34.1 ± 1.7 46.2 ± 5.7

10.6 ± 0.7 11.2 ± 0.6

82.8 ± 4.8 55*

45 min Control CPB

7.47 ± 0.04 7.38 ± 0.04

312 ± 19 527 ± 48*

32.5 ± 2.6 34.6 ± 4.7

10.4 ± 0.5 Il.l ± 0.7

82.2 ± 5.0 55*

75 min Control CPB

7.47 ± 0.03 7.43 ± 0.03

331 ± 24 451 ± 26*

34.6 ± 2.8 37.3 ± 1.5

10.5 ± 0.5 11.3 ± 0.7

78.7 ± 7.7 55*

105 min Control CPB

7.48 ± 0.03 7.42 ± 0.03

312 ± 25 399 ± 45

32.9 ± 1.4 29.6 ± 2.1

10.4 ± 0.7 11.8 ± 0.7

80.8 ± 8.5 55*

CPR. Cardiopulmonary bypass; (J 5. 45. 75. 105 min, specific time periods after administration of heparin (control) or beginning CPB (experimental) (N = 6 for each group); Paolo arterial oxygen tension; Pacos, arterial carbon dioxide tension; MAP, mean arterial blood pressure. Values are expressed as mean ± standard error of the mean. • p < 0.05 versus control group.

Materials and methods Surgical manipulations. Twelve piglets aged 4 weeks (6 to 8 kg) were anesthetized with pentobarbital (Nembutal) (40 mg/kg intra peritoneally). Tracheostomy was performed, and the piglets' lungs were ventilated with 40% oxygen with use of a Harvard ventilator. Respiratory rate and tidal volume were adjusted to maintain arterial carbon dioxide tension between 25 and 35 mm Hg. Catheters were placed in each femoral artery for pressure monitoring and blood sampling and in one femoral vein for intravenous fluid administration. Additional doses of pentobarbital (50 mg intravenously) were given every 30 to 60 minutes as required. Blood pressure was monitored continuously with a Gould P23 ID pressure transducer and a Gould amplifier and recorder (Viggo Spectramed Inc., Critical Care Division, Oxnard, Calif.). A median sternotomy was performed, and each animal received heparin (300 U /kg intravenously). The six control piglets had no further surgical intervention; rectal temperature was maintained at 38° ± 0.5° C by a warming blanket and arterial blood gases, and hemoglobin was determined 15 minutes after administration of heparin and every 30 minutes thereafter. CPB. In the six experimental animals CPB was initiated via cannulas in the aortic root and right atrium. A pulmonary artery vent was employed. The bypass circuit included a Sarns roller pump (Sarns, Inc., Ann Arbor, Mich.), a Bentley-IO Plus bubble oxygenator (Baxter Healthcare Corp., Bentley Laboratories Division, Irvine, Calif.), and a 40 ttm in-line arterial filter. The prime consisted of 1500 to 2500 ml fresh homologous blood to which 10 mEq of sodium bicarbonate was added. A blood prime was used to avoid excessive hemodilution on bypass; mean hemoglobin pre-CPB was 10.6 mg/dl. Mean arterial blood pressure (MAP) was maintained at 55 mm Hg to simulate CPB in human beings. Piglets were cooled to 28° C (rectal) and maintained at this temperature for the first 90 minutes of CPB.

During the last 30 minutes they were warmed to 38° C. Blood gases and hemoglobin were measured 15 minutes after institution of CPB and every 30 minutes thereafter. Gas inflow to the oxygenator was adjusted to maintain arterial carbon dioxide tension between 25 and 35 mm Hg. Arterial oxygen tension during CPB was 400 to 500 mm Hg. Regional transfer coefficient for AlB. Two hours after heparinization (control) or beginning of CPB (experimental), 200 /lCi of AlB (molecular weight 104, specific gravity 40 to 60 mCi/mmol; Dupont-New England Nuclear Products, Boston, Mass.) was injected into the femoral vein (control piglets) or into the oxygenator reservoir (experimental piglets). Injection was given for 200 seconds by a Harvard infusion pump. Arterial blood samples (1 ml) were drawn every 30 seconds for the first 6 minutes after AlB injection and every minute thereafter until the end of the experiment. Fifteen minutes after AlB infusion, the piglets were killed by bolus injection of potassium chloride into the aortic root (experimental piglets) or left ventricle (control piglets). Both internal jugular veinswere transected, and the brain was flushed of blood by gravity infusion of normal saline (l L) through each carotid artery. The brain was removed, and 250 mg tissue samples were dissected from the hippocampus, caudate nucleus, spinal cord, medulla, pons, midbrain, and diencephalon (two samples each region), superior colliculus and pituitary gland (one sample each), and cerebellum (four samples). Eight cortical samples were obtained from each of the primary supply regions of the anterior, middle, and posterior cerebral arteries and from the anterior-middle and posteriormiddle cerebral artery watershed regions. All samples were placed in glass vials.Two milliliters of dihydroxyacetone (Protosol) (Dupont-New England Nuclear Products, Boston, Mass.) were added to each vial to dissolve the tissue sample, and the vials were placed in a 50° C water bath overnight. Ten milliliters of Biocount (Research Products International Corp., Mt.

1 112

The Journal of Thoracic and Cardiovascular Surgery

Gillinov et al.

35000

30000

J::

.2 ...,

...,...J::


111

-

•

Control

\l

CPB

~

:;:l

0

8c"8 ull)

III

25000

20000

Po.

_"0

<~


8

15000

'"
0:

10000

5000

0

0

10

5

15

Time (minutes) Fig. 1. Plasma AIB concentration for control and CPB piglets (values expressed as mean ± standard error of the mean). Prospect, 111.) were used for the scintillation cocktail on the solubilized tissue and 50 ILl aliquots of plasma. Glacial acetic acid (0.05 ml) was added to neutralize the solution and minimize chemiluminescence. Samples were kept at room temperature for 24 hours before counting beta emissions(LS 3901, Beckman Instruments, Inc., Schiller Park, 111.). Disintegrations per minute were corrected for quench and background by external standards method. Using the equation of Ohno, Pettigrew, and Rapoport,'? as modified by Blasberg, Fenstermacher, and Patlak,14the transfer coefficient Ki (ml/gm/rnin) for AIB was calculated as follows: Ki =

14C(T)

~ j14C(P)dt In this equation, 14C(T) is the concentration of AIB in the tissue sample at the end of the experiment (disintegrations per minute per gram) and 15 fI4C(p)dt is the integrated arterial plasma concentration of AIB during the IS-minute sampling time (minutes X disintegrations per minute per milliliter). Ki therefore represents a quantitative indicator of unidirectional transfer of AIB from blood to brain. Statistical analysis. All values are expressed as mean ± standard error of the mean. Arterial blood gas and hemoglobin values were analyzed using analysis of variance for repeated measures. Student's t test with the Bonferroni correction was used when differences were identified. Ki values were analyzed using a two-way analysis of variance and the Newman-Keuls

test. Ki values for individual brain regions were compared between groups with use of Student's t test. Animal care. All animals received humane care in compliance with the "Principles of Laboratory Animal Care" formulated by the National Society for Medical Research and the "Guide for the Care and Use of Laboratory Animals" prepared by the National Academy of Sciences and published by the National Institutes of Health (NIH Publication No. 86-23, revised 1985).

Results Table I shows arterial blood gases, MAP, and hemoglobin values in control piglets and in those undergoing CPR Statistically significant differences in arterial oxygen tension occurred 45 and 75 minutes after beginning CPB or heparin administration, but these differences were probably of no physiologic importance. MAP in piglets undergoing CPB was maintained at 55 mm Hg, which was significantly different from control piglets, in which MAP was generally about 80 mm Hg. There were no statistically significant differences in pH or arterial carbon dioxide tension. Fig. I depicts plasma AlB concentration during the final IS minutes of the experiment. In both groups there was a rapid initial climb in plasma AlB concentration followed by a gradual decline to near steady-state levels.

Volume 104 Number 4 October 1992

CPB and BBB

0.006

0.10

o

E;Sj

Control CPB

0.09

0005

0.08 0.07

0.004

?

a

<,

""

<,

"8

0003

:2 0002

0.001

1 113

~

~

"

JJ

<,

0.04

:2

a ""

]

0.02

~

0.01 //

CER MED PONS MID COll DIE

0.05

0.03

0.00

//

SC

?

<,

~ ~ ~

0.06

HIP CAUD ACA MCA PCA AMW PMW

PIT

Brain Region

Fig. 2. AlB transfer coefficient (Ki) for each brain regionwith p values: spinalcord (SC, p = 0.48 for controlversus pigletsundergoingCPB); cerebellum(CER,p = 0.22); medulla (MED,p = 0.7I); pons(PONS,p = 0.49); midbrain (MID, p = 0.86); superior colliculus (COLL, p = 0.35); diencephalon (DIE, p = 0.50); hippocampus (HIP, p = 0.15); caudate (CAUD, p = 0.16); anterior cerebral artery (ACA, p = 0.85); middle cerebral artery (MCA, p = 0.42); posteriorcerebralartery (PCA,p = 0.37); anterior-middle cerebralarteries watershed (AMW,p = 0.37); posterior-middle cerebral arteries watershed (PMW, p = 0.49); and pituitary (PIT, p = 0.07). Note separate scale at right for pituitary gland. All values expressed as mean ± standard error of the mean.

Because piglets undergoing CPB had a greater total blood volume and the AlB was injected into the oxygenator rather than intravenously, the peak AlB concentration was lower and occurred somewhat later than in control piglets. Mean Ki values for 14 brain regions are shown in Fig. 2. Excluding pituitary Ki, Ki values ranged from 0.0023 to 0.0041 ml/gm/rnin. There was no significant difference between control and CPB Ki values for any brain region. Two-way analysis of variance demonstrated that spinal cord Ki was significantly greater than pons, medulla, superior colliculus, hippocampus, and caudate nucleus in both groups of piglets (p < 0.05 for each region versus spinal cord). Mean pituitary Ki was 0.077 ± 0.012 for control piglets and 0.048 ± 0.008 for piglets undergoingCPB (p = 0.07). Pituitary Ki was 10-fold to 20-fold greater than in other brain regions. Discussion These experiments show that 2 hours of hypothermic CPB does not disrupt the piglet BBB. Two hours of CPB was chosen as representative of routine cardiac operations. Transfer coefficient values were similar in all brain

regions except the pituitary, where Ki was high, reflecting a lack of BBB in the pituitary gland. Our values for Ki are in agreement with those obtained by other investigators.!': 16-18,20 Minor regional variability in Ki has been noted previouslyI6-18, 20 and is of similar magnitude to that found in this study. Statistical analysis revealed no differences in Ki between control piglets and those undergoing CPB in any brain region. These studies were undertaken to determine if BBB disruption occurred during CPB and might explain some neurologic deficits after cardiac operations. Although mortality from cardiac surgical procedures has decreased, neurologic complications still occur with significant frequency. The prevalence of overt stroke after coronary artery bypass graft is 1% to 3%,1-4,7,8,21 and neurologic events are the second most common cause of death after coronary artery bypass graft-' Subtle cognitive and psychologic deficits occur in up to 60% of patients after CPB. 1, 5, 6 Neurologic events after CPB have traditionally been attributed to cerebral hypoperfusion or emboli,1,5,7-10 but controversy still attends their cause; other potential mechanisms must be considered. While massive strokes are often caused by recognized

The Journal of Thoracic and Cardiovascular Surgery

I I I 4 Gillinov et al.

periods of arterial hypotension or large emboli, subtle temporary deficits are more difficult to explain. CPB itself causes "postperfusion syndrome," I I characterized by increased capillary permeability throughout the body. I 1-13 Until now the effects of this generalized inflammatory state on BBB function had not been examined. The brain is protected by the BBB, which consistsof tight junction between capillary endothelial cells and probably also involves active transport systems in these cells.!" Hypertension 15, 22, 23 and ischemia'" can disrupt the BBB. Disruption of the BBB after CPB might cause neurologic dysfunction by development of cerebral edema. In addition, a dysfunctional BBB might allow endogenous and exogenous amines access to cerebral vascular smooth muscle and neurons where they could affect cerebral blood flow and metabolism. These experiments demonstrate that disruption of the BBB cannot be implicated as a cause of neurologic dysfunction after 2 hours of CPB. Alternatively, there exist other explanations for the lack of detectable BBB dysfunction after 2 hours of CPB. First, it is possible that the AlB technique lacks the sensitivity to demonstrate BBB disruption in this model. This is unlikely, since other investigators have employed this technique successfully (including studies of BBB function during cardiopulmonary resuscitation) with use of small'S 16-18 and large'? animals. In this study sensitivity of the technique was demonstrated by finding that the pituitary gland, which has no BBB, had Ki values 10 to 20 times greater than other brain regions. A second possibility is that flushing the brain with normal saline at the experiment's conclusion might wash out AlB that had entered brain cells, thereby underestimating Ki. Given that Ki values derived from these experiments were of the same magnitude as those reported by others.P: 16-18,20 this explanation is unlikely. AlB has a low rate of transport across normal brain capillaries and is rapidly concentrated by brain cells once it crosses the capillary membrane. 14,24 Because AlB is rapidly taken up from the extracellular fluid and essentially trapped by neurons, brain-to-blood backflux is minimized. 14, 24, 25 The high Ki values for the pituitary gland in each experiment also demonstrated that significant washout of AlB did not occur. Finally, it is possible that longer periods of CPB might disrupt the BBB. In this experiment a period of 2 hours was chosen to simulate most adult cardiac operations. Further experiments are required to investigate the effects on the BBB of longer periods of CPB. With use of a model that closely simulates human CPB, these studies show no disruption of the BBB in piglets. It is likely that human beings undergoing 2 hours of

moderately hypothermic CPB have a similarly intact BBB. We wish to thank Mr. Louis Jackson for expert technical assistance, Mr. Joseph DiNatale for assistance with statistical analysis, and Mrs. Candace Berryman and Mrs. Lori Garrison for preparation of the manuscript. REFERENCES I. Govier AV. Central nervous system complications after cardiopulmonary bypass. In: Tinker JH, ed. Cardiopulmonary bypass: current concepts and controversies. Philadelphia: WB Saunders, 1989:41-68. 2. Cosgrove DM, Loop FD, Lytle BW, et al. Primary myocardial revascularization: trends in surgical mortality. J THORAC CARDIOVASC SURG 1984;88:673-84. 3. Gardner TJ, Horneffer PJ, Manolio TA, HoffSJ, Pearson TA. Major stroke after coronary bypass surgery: changing magnitude of the problem. J Vase Surg 1986;3:684-94. 4. Loop FD, Lytle BW, Gill CC, Golding LA, Cosgrove DM, Taylor Pc. Trends in selection and results of coronary artery reoperations, Ann Thorac Surg 1983;36:380-8. 5. Savageau JA, Stanton B, Jenkins D, Klein MD. Neuropsychological dysfunction following cardiac operation. I. Early assessment. J THORAC CARDIOVASC SURG 1982;84:58594. 6. Savageau JA, Stanton B, Jenkins D, Frater R W. Ne,uropsychological dysfunction following elective cardiac operation. II. A six-month reassessment. J THORAC CARDIOVASC SURG 1982;84:595-600. 7. Shaw PJ, Bates D, Cartlidge NE, Heaviside D, Julian DG, Shaw DA. Early neurological complications of coronary artery bypass surgery. Br Med J 1985;291:1384-6. 8. Abuerg T, Kihlgren M. Cerebral protection during open heart surgery. Thorax 1977;32:525-33. 9. Solis RT, Kennedy PS, Beall AC, Noon GP, DeBakey MA. Cardiopulmonary bypass: microembolization and platelet aggregation. Circulation 1975;52:103-8. 10. Bojar RM, Najafi H, DeLaria GA, Serry C, Goldin MD. Neurological complication of coronary revascularization. Ann Thorac Surg 1983;36:427-32. 11. Kirklin JW. The postperfusion syndrome: inflammation and the damaging effects of cardiopulmonary bypass. In: Tinker JH, ed. Cardiopulmonary bypass: current concepts and controversies. Philadelphia: WB Saunders, 1989:13146. 12. Smith EE, Naftel DC, Blackstone EH, Kirklin JW. Microvascular permeability after cardiopulmonary bypass. J THORAC CARDIOVASC SURG 1987;94:225-33. 13. Bando K, Pillai R, Cameron DE, et al. Leukocyte depletion ameliorates free radical-mediated lung injury after cardiopulmonary bypass. J THORAC CARDIOVASC SURG 1990;99:873-7. 14. Blasberg RG, Fenstermacher JD, Patlak CS. Transport of o-aminoisobutyric acid across brain capillary and cellular membranes. J Cereb Blood Flow Metab 1983;3:8-32. 15. Ellison MD, PovlishockJT, Hayes RL. Examination of the

Volume 104 Number 4 October 1992

16.

17.

18.

19.

20.

blood-to-brain transfer of e-aminoisobutyric acid and horseradish peroxidase: regional alterations in blood-brain barrier function following acute hypertension. J Cereb Blood Flow Metab 1986;6:471-80. Tyson GW, Teasdale GM, Graham or, McCulloch J. Cerebrovascular permeability following M CA occlusion in the rat. J Neurosurg 1982;57:186-96. Todd NV, Picozzi P, Crockard RD, Ross-Russell RW. Duration of ischemia influences the development and resolution of ischemic brain edema. Stroke 1986;17:466-71. Dobbin JH, Crockard A, Ross-Russell RW. Transient blood-brain barrier permeability following profound temporary global ischemia: an experimental study using 14CAlB. J Cereb Blood Flow Metab 1989;9:71-8. Ohno K, Pettigrew KD, Rapoport S1. Lower limits of cerebrovascular permeability to nonelectrolytes in the conscious rat. Am J PhysioI1978;235:H299-307. Schleien CL, Koehler RC, Shaffner DH, Traystman RJ. Blood-brain barrier integrity during cardiopulmonary resuscitation in dogs. Stroke 1990;21:1185-91.

CPB and BBB

1 115

21. Gardner TJ, Hornoffer PJ, et al. Stroke following coronary artery bypass grafting: a ten-year study. Ann Thorac Surg 1985;40:574-81. 22. Mayhan WG, Faraci FM, Heistad DD. Disruption of the blood-brain barrier in cerebrum and brain stem during acute hypertension. Am J PhysioI1986;251:HI71-5. 23. Baumbach GL, Heistad DD. Heterogeneity of brain blood flow and permeability during acute hypertension. Am J PhysioI1985;249:H629-7. 24. Blasberg R, Lajtha A. Substrate specificity of steady-state amino acid transport in mouse brain slices. Arch Biochem Biophys 1965;112:361-77. 25. Blasberg RG, Patlak CS, Fenstermacher JD. Selection of experimental conditions for the accurate determination of blood-brain transfer constants from single-time experiments: a theoretical analysis. J Cereb Blood Flow Metab 1983;3:215-25.