- Email: [email protected]
Regional blood flow distribution during extracorporeal membrane oxygenation in rabbits
J
THORAC CARDIOVASC SURG
1989;98: 1138-43
Regional blood flow distribution during extracorporeal membrane oxygenation in rabbits To study regional blood distribution during extracorporeal membrane oxygenation, we stabilized three groups of five rabbits each (3 to 5 kg) on venoarterial bypass at a flow rate of 30 m1jkgjmin. Albumin aggregates (15 to 30 ~m) labeled with technetium 99m were injected into tbe left ventricle during bypass (ventricle), the perfusion cannula during bypass (cannula), and the left ventricle with no bypass (control). Animals were put to death, organs were removed, and the percent distribution was determined with a gamma camera. The Student Newman-Keuls test was used for statistical comparisons. Distribution to both the heart and brain in the cannula group were decreased from control by 55 % and 35%, respectively. Distribution to the brain in the ventricle group was also decreased from control by 39 %. Intestinal distribution was elevated above control in the ventricle group by 37 % , whereas musculoskeletal distribution was elevated 33 % above control in the cannula group. No. significant changes were noted for the kidneys, stomach, or liver. These data suggest that overall perfusion of some vital organs may be significantly reduced during low-flow extracorporeal membrane oxygenation, specifically in the case of the heart and brain, which may be deprived of oxygenated blood.
Todd T. Nowlen, MS,a Steven O. Salley, PhD,b Grant C. Whittlesey, CCP,c Sourav K. Kundu, MS,b Nancy A. Maniaci, CNMT,d Raymond L. Henry, PhD,a and Michael D. Klein, MD,c. e Detroit. Mich.
Extracorporeal membrane oxygenation (ECMO) has become common treatment for newborn infants with respiratory failure unresponsive to less invasive forms of therapy. I Recent studies indicate it may also have a place in the treatment of cardiac failure- 3 and respiratory failure in adults." Although the physiology of extracorporeal circulation (ECC) as applied in cardiac surgery has been studied, the applicability of these studies to ECMO is not certain because ECMO is different in several important ways. During cardiopulmonary bypass in the operating room, no vessels are ligated, drainage of both the inferior and superior venae cavae is captured, and the perfusion
From the Departments of Physiology," Chemical Engineering," Surgery," and Radiology," School of Medicine and College of Engineering, Wayne State University and the Children's Hospital of Michigan," Detroit, Mich. Source of funding: Children's Hospital of Michigan, Detroit, Mich. Received for publication Aug. 1, 1988. Accepted for publication March 15, 1989. Address for reprints: Michael D. Klein, MD, Department of Surgery, Children's Hospital of Michigan, 3901 Beaubien Blvd., Detroit, MI 48201.
12/1/13090
1 1 38
cannula is in the aortic arch or retrograde from the femoral artery. The myocardium is protected with intermittent doses of cold cardioplegic solutions injected into the aortic root, and often hypothermia is applied. During ECMO, however, the internal jugular vein and carotid artery on the right side are permanently ligated distal to the cannulas. Only part of the venous return to the heart is captured with a venous cannula in the right atrium. The perfusion cannula is in the innominate artery, and no special protection can be applied to the heart or coronary circulation. For this reason we studied regional blood flow distribution in rabbits on ECMO by using radiolabeled macroaggregated albumin (MAA). Materials and methods New Zealand White rabbits weighing 3.0 to 4.5 kg were anesthetized with intramuscular injections ofketamine 35 mg/ kg and xylazine 5 rug/kg. Anesthesia was maintained with ketamine and xylazine given into a marginal ear vein as needed. The rabbits were intubated through a tracheostomy and their lungs ventilated with a mixture of air, oxygen, or carbogen (95% oxygen and 5% carbon dioxide) on a volume-cycled ventilator to maintain normal blood gases. Heparin (300 units/kg) was administered before cannulation, which generally maintained the activated clotting time at greater than 500 seconds for the duration of the experiment. An additional bolus of 300 units of
Volume 98 Number 6
Regional blood flow distribution during ECMO
December 1989
I I 39
o
D
HEPARIN
INFUSION
MEMBRANE OXYGENATOR
ROllER PUMP
Fig. I. ECMO circuit diagram showing route of venous blood from right atrium through roller pump, oxygenator, and into right carotid artery. heparin was administered if the activated clotting time dropped below 500 seconds. The left femoral artery was cannulated for blood pressure monitoring and blood sampling. The right common carotid artery and external jugular vein were isolated for cannulation. The ECMOcircuit used in this experiment (Fig. I) consisted of a 0.4 m2 SciMed silicone membrane oxygenator (SciMed Life Systems, Inc., Minneapolis, Minn.) and a Sarns SIOKII roller pump (Sarns, Inc., Ann Arbor, Mich.). A syringe positioned between the venous return line and the pump acted as a reservoir. The tubing (1f4-inch polyvinylchloride) was kept as short as possible and the entire circuit was maintained at the same level as the rabbit. The rabbit was warmed by a lamp or pad to help maintain normal body temperature. The circuit was primed and debubbled with normal saline containing heparin 40 units/nil, which was subsequently replaced with anticoagulated rabbit blood (Pel-Freez, Rogers, Ark.) or donor blood. Flow rates during ECMO were maintained at 30 nil/kg/min. This ECMO procedure was conducted much as is done in patients. Crystalloid and blood were administered as needed to maintain blood pressure. The membrane lung was ventilated with mixtures of air, oxygen, and carbogen as needed to maintain normal arterial blood gases. . Three groups of five animals each were studied. Group I (control) had a 19- or 22-gauge polyethylene cannula passed down the right common carotid artery into the left ventricle (LV) with distal ligation of the artery. MAA labeled with technetium 9mTc-MAA; Pulmonite, Billerica, Mass.), more than 90% being 10 to 90 ,urn,average size IS to 30 ,urn, were injected into this cannula. Flow distribution in this group represented normal distribution of blood ejected from the LV in animals not on ECMO. Group 2 (cannula) had a 10F or 12F chest tube passed down the external jugular vein into the right atrium (or the inferior vena cavajust inferior to the right atrium) for venous drainage. An 8F chest tube cut off above the side holes was passed down the carotid artery to a level just cranial to the right
e
and left common carotid/right subclavian junction (Fig, 2). In this group injections of the 99mTc-MAA were made into the perfusion cannula to represent blood flow delivered by the circuit during ECMO. Cannula placement in group 3 (ventricle) was identical to that in group 2, except that a Y connector was placed on the arterial cannula so that a 22-gauge polyethylene cannula could be threaded down the center into the LV. 99mTc_ MAA was injected into the LV cannula to measure blood flow distribution from the heart during ECMO. Cannula placement was verified in all groups by roentgenograms and pressure waveforms. After cannula placement and the initiation of ECMO, the rabbit's physiologic parameters (blood pressure, heart rate, and arterial oxygen and carbon dioxide tension) were stabilized for IS minutes. Then approximately 5 mCi of 99mTc_MAA suspended in 4 ml normal saline was injected. After IS minutes the animals were put to death by intravenous pentobarbital 120 mg/kg followed by bilateral pneumothoraces. Placement of arterial and venous cannulas and of the LV injection cannula was again verified, and all major organs were removed for counting. The isolated organs and the remaining musculoskeletal tissues (all tissue remaining after removal of the listed organs) were counted individually on a Siemens ZLC 370 gamma camera (Siemens Gammasonics Inc., Nuclear Medicine Division, Des Plains, Ill.). Because the diameter of the 99mTc_MAA is greater than that of the capillary, injections into the arterial system caused the aggregates to be trapped within the first capillary bed they encountered. Relative flow to a specific organ was determined as the percent of the total number of counts in the animal. Data were analyzed by the Student N ewman-Keuls test with p < 0.05 considered significant. 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.
The Journal of
1 140
Thoracic and Cardiovascular Surgery
Now/en et a/.
LEFT
. / COMMON CAROTID -LEFT SUBCLAVIAN
ARTERY
"-
LEFT
CRANIAl Vl:NA CAVA
ferences in the distribution of the blood originating in the ECMO circuit when compared with normal LV distribution. A significant reduction in the percentage of blood flow that perfuses the heart (-55%) and brain (total, left hemisphere, and hindbrain: -35%, -33%, and -43%, respectively) is noted. In addition, the percent that perfuses the musculoskeletal tissues is increased (+ 33%). Similarly, significant differences are noted in the distribution of the blood originating in the LV during ECMO when compared with normal LV distribution. Reductions in percent of blood flow to the brain (total, left hemisphere, and hindbrain: -39%, -40%, and -42%, respectively), both lungs (-46%), and spleen (-48%) are observed, whereas there is an increased fraction (+37%) to the intestines. Comparison of LV flow during ECM0 to cannula flow during ECMO shows that a significantly greater fraction of the LV flow perfuses the heart ( +77%) than that from the cannula, and the fraction of LV flow to the intestines is also greater (+56%). No significant differences in flow were noted between the left and right kidneys, lungs, or cerebral hemispheres within any group. Discussion
Fig. 2. Gross anatomy of aortic root and great vessels in rabbit,showing sitesof arterial and venous cannulationfor ECMO.
Results Physiologic parameters in the three groups of rabbits are presented in Table I and statistically significant differences are noted. The hematocrit value and volume of oxygen bound to hemoglobin (percent) were lower in the ECMO groups because of hemodilution despite the administration of donor blood. In addition, arterial oxygen tension was higher in the ventricle group because of the external oxygenation of the bypass blood, whereas pH in the cannula group was significantly lower than control. None of the other measured parameters are remarkably different in any group and virtually no significant differences are noted between ECMO groups. Despite these differences, none of the parameters for any of the groups are outside clinically acceptable limits. Table II reports the animal weights, durations of anesthesia and bypass, and flow rates. No significant differences were noted between any groups for these measures. Flow distribution data for all groups are reported in Table III. Organ blood flow to the excised organ is reported as the percent of the total counts injected, and statistically significant differences between groups are noted. The results indicate that there are significant dif-
ECC is common practice in many cardiac surgical procedures. Much work has been done to document the physiologic changes seen during ECC. Early studies of the redistribution of blood flow suggested that while total flow remained constant, an increase in blood pressure would redirect blood flow to organs more responsive to perfusion pressure and less responsive to arteriolar tone.' The two most important organs, the brain and heart, were thus preserved whereas organs responsive to increased arteriolar tone such as the skin and skeletal muscles had a decreased perfusion. An increase in blood pressure, however, is not seen in all ECC procedures. Other early work suggested that redistribution of blood during ECC was due not only to changes in pulse pressure and nonpulsatile perfusion but also to other factors.t" including anesthesia, peripheral emboli, hemodilution, and the effects on blood elements. Much emphasis in recent years has been placed on the evaluation of cerebral blood flow (CBF) during ECC. Both decreased 6• 8- JO and increased'v '? CBF has been documented. Impaired or abolished autoregulation ofCBF during ECCI J. 13. 14 has also been demonstrated while CBF sensitivity to hypercapnia and hemodilution is still maintained. 14. 15 Several studies in lambs, rats, and rabbits have shown that ligation of the common carotid artery alone does not alter distribution of blood between the left and right cerebral hemispheres.l'v'f Each study used a technique comparable with that used in this experiment; that is, in-
Volume 98 Number 6
Regional blood flow distribution during EeMO
December 1989
1 14 1
Table I. Physiologic values during blood flow distribution studies Parameter Heart rate (beats/min) Systolic BP (mm Hg) Diastolic BP (mm Hg) Hematocrit (vol %) Vol. O 2 bound to Hgb (%) Arterial pH Pao2 (mm Hg) 02 sat. (%) Paco2 (mm Hg) HC0 3 (mE/L)
Group I: Control
164 ± 79 ± 54 ± 39 ± 16.6 ± 7.46 ± 110 ± 98 ± 34 ± 24 ±
13 7 13 2 1.3
0.03 11 0.5 2 2
Group 2: Cannula
156 ± 68 ± 47 ± 30 ± 13.3 ± 7.32 ± 151 ± 98 ± 38 ± 19 ±
30 30 18 5* 1.9* 0.12* 45 2 9 4
Group 3: Ventricle
130 ± 20 67 ± 13 51 ± 8 32 ± 5* 13.2 ± 1.2* 7.37 ± 0.03 190±36t 99 ± 0.3 34 ± 6 21 ± 4
Values represent mean ± standard deviation (N = 5). BP. Blood pressure; Hgb, hemoglobin; Pao-, arterial oxygen tension; Paco-, arterial carbon dioxide tension: HeO l , bicarbonate radical. of'
tf'
< 0.05 < 0.01
when compared with control. when compared with control.
Table II. Experimental variables during blood flow distribution studies Parameter
Group I: Control
Weight (kg) Anesthesia duration (min) Bypass duration (min) Flow (ml/kg/min)
3.46 ± 0.20 200 ± 55 N/A N/A
Group 2: Cannula
4.01 ± 249 ± 121 ± 29.4 ±
0.50 86 56 5.4
Group 3: Ventricle
3.90 ± 229 ± 126 ± 30.2 ±
0.82 58 51 0.2
Values represent mean ± standard deviation (N = 5). Nj A, Not applicable.
jecting microspheres into the LV via a catheter placed in the ligated common carotid artery. Also using a similar technique in piglets, Laptook, Stonestreet, and William19 demonstrated that ligation of the common carotid artery does not alter total blood flow to the brain. The effects of ligation of both the carotid artery and the jugular vein, however, have not been studied previously. Data from this study indicate that regional blood flow distribution during ECMO is not the same as when the entire cardiac output is delivered from the LV (Table III). Even the regional distribution of blood from the output of the LV is altered when blood is entering the aorta from the perfusion cannula. Distribution of 99mTc_MAA to the total brain was decreased during ECMO as compared with control when it originated in either the LV or the perfusion cannula. In the rabbit, the innominate artery branches into the left and right common carotid arteries and the right subclavian artery (Fig. 2). The hindbrain is supplied mainly by the vertebral arteries.P If flow from the ECMO circuit travels preferentially cranially via the carotids and caudally via the descending aorta, this could explain the decreased hindbrain flow in both bypass groups. The subclavian arteries may not be fed by the ECMO circuit, or perhaps the right subclavian might be supplied by the ECMO circuit while the left is not. Because of the rabbit anatomy, the possibility also exists
that blood perfusing the brain may come predominantly from the ECMO circuit while total flow is still less than control. Hemodilution has been shown to cause an increased CBF, probably related to a decreased viscosity.2l,22 An increase in CBF could compensate for the diminished oxygen-carrying capacity.F" 24 Despite these phenomena, in this experiment, in addition to a decreased relative blood flow to the brain, there resulted a decreased relative total oxygen delivery because of the significantly lowered hematocrit value. The myocardium may also have a lowered total oxygen delivery during ECMO. The decreased percentage of counts in the heart when the 99mTc-MAA originated in the ECMO circuit, as compared with when it originated in the LV in the control group, suggested that the myocardium during ECMO is not well perfused with blood from the ECMO circuit. This is supported by the fact that 99mTc_MAA distribution to the heart from the perfusion cannula during ECMO was also significantly less than that exiting the LV during ECMO. Rabbit great vessel anatomy is similar enough to the human anatomy to conclude that perfusion of the heart in the human via the low-flow ECMO circuit might also be decreased. Data from this study suggest that blood from the ECMO circuit may not pass retrogradely down the ascending aorta in sufficient amounts. This hypothesis is supported by
The Journal of Thoracic and Cardiovascular Surgery
Nowlen et al.
1 1 42
Table III. Regional blood flow distribution in controls and rabbits on ECMO (reported as percent of total counts) Organ Left brain Right brain Hindbrain Total brain Heart Left lung Right lung Left kidney Right kidney Stomach Intestine Liver Spleen Musculoskeletal (carcass)
Group 1: Control
3.37 ± 3.05 ± 3.32 ± 9.75 ± 3.95 ± 2.93 ± 3.05 ± 8.94 ± 8.68 ± 4.86 ± 10.56 ± 7.67 ± 3.49 ± 35.93 ±
1.03 1.00 1.17 3.00 1.11 0.93 0.98 1.76 1.6I 0.56 1.74 2.43 1.30 4.66
Group 2: Cannula
2.27 ± 2.20 ± 1.89 ± 6.36 ± 1.79 ± 2.08 ± 2.05 ± 6.53 ± 6.22 ± 7.81 ± 9.29 ± 7.53 ± 2.52 ± 47.83 ±
0.46* 0.44 0.25* 1.13* 0.39t:j: 0.50 0.30* 1.93 2.07 4.57 I.83§ 2.70 . 0.48 8.08*
Group 3: Ventricle
2.01 ± 0.57* 1.96 ± 0.58 1.92 ± 0.70* 5.89 ± 1.83* 3.17±0.78 1.59 ± 0.66* 1.65 ± 0.71* 7.91 ± 1.83 7.68 ± 1.83 7.12 ± 3.09 14.50 ± 2.01* 8.15 ± 1.16 1.81 ± 0.98* 40.54 ± 5.17
Values represent mean ± standard deviation (N : 5) .
• p < 0.05 when compared with control. tp
< 0.01
when compared with control.
:j:p < 0.05 when cannula is compared with ventricle.
§p
< 0.01
when cannula is compared with ventricle.
Seeker-Walker and associates.P who demonstrated during venoarterial bypass that even when 85% of the systemic output originated in the bypass circuit only 25% of the coronary artery flow originated in the bypass circuit. Of special significance during clinical ECMO is the fact that cannula blood is well oxygenated and alkalotic, whereas LV blood may be poorly oxygenated and acidotic. Thus the relative lack of cannula flow to the heart may be even more physiologically significant. An increased relative blood flow was noted to the intestine in the ventricle group as compared with the control group (Table III). Although the mechanism for the increased flow to the splanchnic regions during ECC is unknown, this phenomenon has been documented by other researchers in various animals.v 8, 21.26, 27 Distributions to the heart and intestines were the only two significant differences noted between the two bypass groups (Table III). This suggests that overall mixing of blood from the perfusion circuit and the LV (with the possible exceptions of distribution to heart and intestine) is adequate. This conclusion is supported by the observation that there were no significant differences in flow between the left and right kidneys, lungs, and cerebral hemispheres within any group. Effects of anesthesia on regional blood flow in rabbits have been reported. 28- 32 Xylazine and ketamine in the same combination used in this study were shown to produce a 30% decrease in blood pressure, a 77% decrease in respiratory rate, and a 28% reduction in heart rate." No studies were found on the regional distribution of cardiac output in the rabbit when ketamine and xylazine were used as the anesthetic. Dhasma and associa tes, 29 however,
have studied the effects of 2, 6, and 18 rug/kg doses of ketamine in rabbits. Cardiac output and blood flow to many tissues (including total brain and heart) remained unchanged. In contrast, Blake and Komer.l? using a 12 rng/kg dose of ketamine in rabbits, demonstrated an increase in arterial pressure, cardiac output, and heart rate with no significant change in total peripheral resistance. These studies suggest that the anesthetic used in this study may not affect cardiac output distribution. In the event that the anesthetic does influence regional blood flow, the effect of ECMO should still be evident, because all animals were anesthetized, and there were no significant differences in the duration of anesthesia among the three groups. The methods used in this study do not allow the determination of absolute blood flows during ECMO. Future experiments should be designed to determine absolute flows. The blood perfusion rate of 30 nil/kg/min used in this study was the maximum we could obtain with this model. This value is less than the usual flows of 100 ml/ kg/min used to fully support the human newborn infant with ECMO, so that conclusions drawn from this study must be viewed with caution. Several conclusions can be drawn from these studies. First, normal regional blood flow distribution in patients during ECMO cannot be assumed. Second, the heart and total brain may be significantly deprived of oxygenated blood during ECMO. Third, adequate mixing of blood pumped by the perfusion circuit and the heart does occur in the rabbit during low-flow ECMO with the possible exceptions of the heart and intestines. If this is indeed the case, then a better argument might be made for veno-
Volume 98 Number 6 December 1989
venous bypass in patients with lung disease alone, and special care should be given in choosing and cannulating patients with cardiac failure. REFERENCES 1. Toomasian JM, Snedecor SM, Cornell RG, Cilley RG, Bartlett RH. National experience with extracorporeal membrane oxygenation for newborn respiratory failure. ASAIO Trans 1988;34:140-7. 2. Redmond CR, Graves ED, Falterman KW, Ochsner JL, Arensman RM. Extracorporeal membrane oxygenation for respiratory and cardiac failure in infants and children. J THORAC CARDIOVASC SURG 1987;93: 199-204. 3. Kanter KR, Pennington DG, Weber TR, Zambie MA, Braun P, Martychenko V. Extracorporeal membrane oxygenation for postoperative cardiac support in children. J THoRAc CARDIOVASC SURG 1987;93:27-35. 4. Iatridis A, Voorhees M, Miller M. Review of ECMO in adults: North American experience. Perfusion 1988;3:3740. 5. Bartlett RH, Gazzaniga AB. Physiology and pathophysiology of extracorporeal circulation. In: Ionescu MI, ed. Techniques in extracorporeal circulation. 2nd ed. London: Butterworth, 1981:5-6. 6. Lees MH, Herr RH, Hill JD, et al. Distribution of systemic blood flow of the rhesus monkey during cardiopulmonary bypass. J THORAC CARDIOVASC SURG 1971;61:570-86. 7. Rudy LW, Heymann MA, Edmunds LH. Distribution of systemic blood flowduring total cardiopulmonary bypass in rhesus monkeys. Surg Forum 1970;21: 149-51. 8. Rudy LW, Heymann MA, Edmunds LH. Distribution of systemic blood flowduring cardiopulmonary bypass. J Appl Physiol 1973;34:194-200. 9. Sorenson HR, Husum B, Waaben J, et al. Brain microvascular function during cardiopulmonary bypass. J THORAC CARDIOVASC SURG 1987;94:727-32. 10. Govier A V, Reves JG, McKay RD, et al. Factors and their influence on regional cerebral blood flow during nonpulsatile cardiopulmonary bypass. Ann Thorac Surg 1984; 38:592-600. II. Santillan GG, Chemnitius JM, Bing RJ. The effect of cardiopulmonary bypass on cerebral blood flow. Brain Res 1985;345: 1-9. 12. Henriksen L, Hjems E, Lindeburgh T. Brain hyperperfusion during cardiac operations. J THORAC CARDIOVASC SURG 1983;86:202-8. 1'3. Lundar T, Lindegaard KF, Froysaker T, Aaslid R, Grip A, Nornes H. Dissociation between cerebral autoregulation and carbon dioxide reactivity during nonpulsatile cardiopulmonary bypass. Ann Thorac Surg 1985;40:582-7. 14. Lundar T, Lindegaard KF, Froysaker T, Aaslid R, Wiberg J, Nornes H. Cerebral perfusion during nonpulsatile cardiopulmonary bypass. Ann Thorac Surg 1985;40: 144-50. 15. Patterson JL. Circulation through the brain. In: Ruch TC, Patton HD, eds. Physiology and biophysics. Philadelphia: Saunders, 1965:953-8. 16. Furzan J, Gabriele G, Wheeler J, Fixler D, Rosenfeld e.
Regional blood flow distribution during EeMO
I 1 43
Regional blood flows in newborn lambs during endotracheal continuous airway pressure and continuous negative pressure breathing. Pediatr Res 1981;15:874-8. 17. Mendell P, Hollenberg N. Cardiac output distribution in the rat: comparison of rubidium and microsphere methods. Am J PhysioI1971;221:1617-20. 18. Neutze JM, Wyler F, Rulolph AM. Use of radioactive rnicrospheres to assess distribution of cardiac output in rabbits. Am J Physiol 1968;215:486-95. 19. Laptook AR, Stonestreet BS, William OH. The effect of carotid artery ligation on brain blood flow in newborn piglets. Brain Res 1983;276:51-4. 20. Baldwin BA, Bell FR. The anatomy of the cerebral circulation of the sheep and ox: the dynamic distribution of the blood supplied by the carotid and vertebral arteries to the cranial regions. J Anat 1963;97:203-15. 21. Utley JR, Wachtel C, Cain RB, Spaw EA, Collins JC, Stephens DB. Effects of hypothermia, hemodilutin, and pump oxygenation on organ water content, blood flow and oxygen delivery, and renal function. Ann Thorac Surg 1981;31:121-33. 22. Gordon RJ, Ravin M, DaicoffGR, Rawitscher RE. Effects of hemodilution on hypotension during cardiopulmonary bypass. Anesth Analg 1975;54:482-8. 23. Tanaka T, Inoue T, Paton Be. Oxygen availability during hypothermic perfusion using diluted blood. Surg Forum 1962;13:138-40. 24. Wright CJ. The effects of severe progressive hemodilution on regional blood flow and oxygen consumption. Surgery 1976;79:299-305. 25. Seeker-Walker JS, Edmonds JF, Spratt EH, Conn A W. The source of coronary perfusion during partial bypass for extracorporeal membrane oxygenation (ECMO). Ann Thorac Surg 1976;21: 138-43. 26. Rudy LW, Boucher JK, Edmunds LH Jr. The effect of deep hypothermia and circulatory arrest on the distribution of systemic blood flow in rhesus monkeys. J THORAC CARDIOVASC SURG 1972;64:706-13. 27. Halley MM, Reemtsma K, Creech 0 Jr. Hemodynamics and metabolism of individual organs during extracorporeal circulation. Surgery 1959;46: 1128-34. 28. Sanford TD, Colby ED. Effect of xylazine and ketamine on blood pressure, heart rate and respiratory rate in rabbits. Lab Anim Sci 1980;30:519-23. 29. Dhasma KM, Saxena PR, Prakash 0, Van Der Zee HT. A study on the influence of ketamine on systemic and regional haemodynamics in conscious rabbits. Arch Int Pharmacodyn 1984;269:323-34. 30. Blake DW, Korner PI. Effects of ketamine and althesin anesthesia on baroreceptor-heart rate reflex and hemodynamics of intact and pontine rabbits. J Auton Nerv Syst 1982;5:145-54. 31. Wyler F, Neutze JM, Rudolph AM. Effects of endotoxin on distribution of cardiac output in unanesthetized rabbits. Am J Physiol 1970;219:246-51. 32. Wyler F. Effect of general anesthesia on distribution of cardiac output and organ blood flow in the rabbit: halothane and chloralose-urethane. J Surg Res 1974;17:381-6.
THORAC CARDIOVASC SURG
1989;98: 1138-43
Regional blood flow distribution during extracorporeal membrane oxygenation in rabbits To study regional blood distribution during extracorporeal membrane oxygenation, we stabilized three groups of five rabbits each (3 to 5 kg) on venoarterial bypass at a flow rate of 30 m1jkgjmin. Albumin aggregates (15 to 30 ~m) labeled with technetium 99m were injected into tbe left ventricle during bypass (ventricle), the perfusion cannula during bypass (cannula), and the left ventricle with no bypass (control). Animals were put to death, organs were removed, and the percent distribution was determined with a gamma camera. The Student Newman-Keuls test was used for statistical comparisons. Distribution to both the heart and brain in the cannula group were decreased from control by 55 % and 35%, respectively. Distribution to the brain in the ventricle group was also decreased from control by 39 %. Intestinal distribution was elevated above control in the ventricle group by 37 % , whereas musculoskeletal distribution was elevated 33 % above control in the cannula group. No. significant changes were noted for the kidneys, stomach, or liver. These data suggest that overall perfusion of some vital organs may be significantly reduced during low-flow extracorporeal membrane oxygenation, specifically in the case of the heart and brain, which may be deprived of oxygenated blood.
Todd T. Nowlen, MS,a Steven O. Salley, PhD,b Grant C. Whittlesey, CCP,c Sourav K. Kundu, MS,b Nancy A. Maniaci, CNMT,d Raymond L. Henry, PhD,a and Michael D. Klein, MD,c. e Detroit. Mich.
Extracorporeal membrane oxygenation (ECMO) has become common treatment for newborn infants with respiratory failure unresponsive to less invasive forms of therapy. I Recent studies indicate it may also have a place in the treatment of cardiac failure- 3 and respiratory failure in adults." Although the physiology of extracorporeal circulation (ECC) as applied in cardiac surgery has been studied, the applicability of these studies to ECMO is not certain because ECMO is different in several important ways. During cardiopulmonary bypass in the operating room, no vessels are ligated, drainage of both the inferior and superior venae cavae is captured, and the perfusion
From the Departments of Physiology," Chemical Engineering," Surgery," and Radiology," School of Medicine and College of Engineering, Wayne State University and the Children's Hospital of Michigan," Detroit, Mich. Source of funding: Children's Hospital of Michigan, Detroit, Mich. Received for publication Aug. 1, 1988. Accepted for publication March 15, 1989. Address for reprints: Michael D. Klein, MD, Department of Surgery, Children's Hospital of Michigan, 3901 Beaubien Blvd., Detroit, MI 48201.
12/1/13090
1 1 38
cannula is in the aortic arch or retrograde from the femoral artery. The myocardium is protected with intermittent doses of cold cardioplegic solutions injected into the aortic root, and often hypothermia is applied. During ECMO, however, the internal jugular vein and carotid artery on the right side are permanently ligated distal to the cannulas. Only part of the venous return to the heart is captured with a venous cannula in the right atrium. The perfusion cannula is in the innominate artery, and no special protection can be applied to the heart or coronary circulation. For this reason we studied regional blood flow distribution in rabbits on ECMO by using radiolabeled macroaggregated albumin (MAA). Materials and methods New Zealand White rabbits weighing 3.0 to 4.5 kg were anesthetized with intramuscular injections ofketamine 35 mg/ kg and xylazine 5 rug/kg. Anesthesia was maintained with ketamine and xylazine given into a marginal ear vein as needed. The rabbits were intubated through a tracheostomy and their lungs ventilated with a mixture of air, oxygen, or carbogen (95% oxygen and 5% carbon dioxide) on a volume-cycled ventilator to maintain normal blood gases. Heparin (300 units/kg) was administered before cannulation, which generally maintained the activated clotting time at greater than 500 seconds for the duration of the experiment. An additional bolus of 300 units of
Volume 98 Number 6
Regional blood flow distribution during ECMO
December 1989
I I 39
o
D
HEPARIN
INFUSION
MEMBRANE OXYGENATOR
ROllER PUMP
Fig. I. ECMO circuit diagram showing route of venous blood from right atrium through roller pump, oxygenator, and into right carotid artery. heparin was administered if the activated clotting time dropped below 500 seconds. The left femoral artery was cannulated for blood pressure monitoring and blood sampling. The right common carotid artery and external jugular vein were isolated for cannulation. The ECMOcircuit used in this experiment (Fig. I) consisted of a 0.4 m2 SciMed silicone membrane oxygenator (SciMed Life Systems, Inc., Minneapolis, Minn.) and a Sarns SIOKII roller pump (Sarns, Inc., Ann Arbor, Mich.). A syringe positioned between the venous return line and the pump acted as a reservoir. The tubing (1f4-inch polyvinylchloride) was kept as short as possible and the entire circuit was maintained at the same level as the rabbit. The rabbit was warmed by a lamp or pad to help maintain normal body temperature. The circuit was primed and debubbled with normal saline containing heparin 40 units/nil, which was subsequently replaced with anticoagulated rabbit blood (Pel-Freez, Rogers, Ark.) or donor blood. Flow rates during ECMO were maintained at 30 nil/kg/min. This ECMO procedure was conducted much as is done in patients. Crystalloid and blood were administered as needed to maintain blood pressure. The membrane lung was ventilated with mixtures of air, oxygen, and carbogen as needed to maintain normal arterial blood gases. . Three groups of five animals each were studied. Group I (control) had a 19- or 22-gauge polyethylene cannula passed down the right common carotid artery into the left ventricle (LV) with distal ligation of the artery. MAA labeled with technetium 9mTc-MAA; Pulmonite, Billerica, Mass.), more than 90% being 10 to 90 ,urn,average size IS to 30 ,urn, were injected into this cannula. Flow distribution in this group represented normal distribution of blood ejected from the LV in animals not on ECMO. Group 2 (cannula) had a 10F or 12F chest tube passed down the external jugular vein into the right atrium (or the inferior vena cavajust inferior to the right atrium) for venous drainage. An 8F chest tube cut off above the side holes was passed down the carotid artery to a level just cranial to the right
e
and left common carotid/right subclavian junction (Fig, 2). In this group injections of the 99mTc-MAA were made into the perfusion cannula to represent blood flow delivered by the circuit during ECMO. Cannula placement in group 3 (ventricle) was identical to that in group 2, except that a Y connector was placed on the arterial cannula so that a 22-gauge polyethylene cannula could be threaded down the center into the LV. 99mTc_ MAA was injected into the LV cannula to measure blood flow distribution from the heart during ECMO. Cannula placement was verified in all groups by roentgenograms and pressure waveforms. After cannula placement and the initiation of ECMO, the rabbit's physiologic parameters (blood pressure, heart rate, and arterial oxygen and carbon dioxide tension) were stabilized for IS minutes. Then approximately 5 mCi of 99mTc_MAA suspended in 4 ml normal saline was injected. After IS minutes the animals were put to death by intravenous pentobarbital 120 mg/kg followed by bilateral pneumothoraces. Placement of arterial and venous cannulas and of the LV injection cannula was again verified, and all major organs were removed for counting. The isolated organs and the remaining musculoskeletal tissues (all tissue remaining after removal of the listed organs) were counted individually on a Siemens ZLC 370 gamma camera (Siemens Gammasonics Inc., Nuclear Medicine Division, Des Plains, Ill.). Because the diameter of the 99mTc_MAA is greater than that of the capillary, injections into the arterial system caused the aggregates to be trapped within the first capillary bed they encountered. Relative flow to a specific organ was determined as the percent of the total number of counts in the animal. Data were analyzed by the Student N ewman-Keuls test with p < 0.05 considered significant. 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.
The Journal of
1 140
Thoracic and Cardiovascular Surgery
Now/en et a/.
LEFT
. / COMMON CAROTID -LEFT SUBCLAVIAN
ARTERY
"-
LEFT
CRANIAl Vl:NA CAVA
ferences in the distribution of the blood originating in the ECMO circuit when compared with normal LV distribution. A significant reduction in the percentage of blood flow that perfuses the heart (-55%) and brain (total, left hemisphere, and hindbrain: -35%, -33%, and -43%, respectively) is noted. In addition, the percent that perfuses the musculoskeletal tissues is increased (+ 33%). Similarly, significant differences are noted in the distribution of the blood originating in the LV during ECMO when compared with normal LV distribution. Reductions in percent of blood flow to the brain (total, left hemisphere, and hindbrain: -39%, -40%, and -42%, respectively), both lungs (-46%), and spleen (-48%) are observed, whereas there is an increased fraction (+37%) to the intestines. Comparison of LV flow during ECM0 to cannula flow during ECMO shows that a significantly greater fraction of the LV flow perfuses the heart ( +77%) than that from the cannula, and the fraction of LV flow to the intestines is also greater (+56%). No significant differences in flow were noted between the left and right kidneys, lungs, or cerebral hemispheres within any group. Discussion
Fig. 2. Gross anatomy of aortic root and great vessels in rabbit,showing sitesof arterial and venous cannulationfor ECMO.
Results Physiologic parameters in the three groups of rabbits are presented in Table I and statistically significant differences are noted. The hematocrit value and volume of oxygen bound to hemoglobin (percent) were lower in the ECMO groups because of hemodilution despite the administration of donor blood. In addition, arterial oxygen tension was higher in the ventricle group because of the external oxygenation of the bypass blood, whereas pH in the cannula group was significantly lower than control. None of the other measured parameters are remarkably different in any group and virtually no significant differences are noted between ECMO groups. Despite these differences, none of the parameters for any of the groups are outside clinically acceptable limits. Table II reports the animal weights, durations of anesthesia and bypass, and flow rates. No significant differences were noted between any groups for these measures. Flow distribution data for all groups are reported in Table III. Organ blood flow to the excised organ is reported as the percent of the total counts injected, and statistically significant differences between groups are noted. The results indicate that there are significant dif-
ECC is common practice in many cardiac surgical procedures. Much work has been done to document the physiologic changes seen during ECC. Early studies of the redistribution of blood flow suggested that while total flow remained constant, an increase in blood pressure would redirect blood flow to organs more responsive to perfusion pressure and less responsive to arteriolar tone.' The two most important organs, the brain and heart, were thus preserved whereas organs responsive to increased arteriolar tone such as the skin and skeletal muscles had a decreased perfusion. An increase in blood pressure, however, is not seen in all ECC procedures. Other early work suggested that redistribution of blood during ECC was due not only to changes in pulse pressure and nonpulsatile perfusion but also to other factors.t" including anesthesia, peripheral emboli, hemodilution, and the effects on blood elements. Much emphasis in recent years has been placed on the evaluation of cerebral blood flow (CBF) during ECC. Both decreased 6• 8- JO and increased'v '? CBF has been documented. Impaired or abolished autoregulation ofCBF during ECCI J. 13. 14 has also been demonstrated while CBF sensitivity to hypercapnia and hemodilution is still maintained. 14. 15 Several studies in lambs, rats, and rabbits have shown that ligation of the common carotid artery alone does not alter distribution of blood between the left and right cerebral hemispheres.l'v'f Each study used a technique comparable with that used in this experiment; that is, in-
Volume 98 Number 6
Regional blood flow distribution during EeMO
December 1989
1 14 1
Table I. Physiologic values during blood flow distribution studies Parameter Heart rate (beats/min) Systolic BP (mm Hg) Diastolic BP (mm Hg) Hematocrit (vol %) Vol. O 2 bound to Hgb (%) Arterial pH Pao2 (mm Hg) 02 sat. (%) Paco2 (mm Hg) HC0 3 (mE/L)
Group I: Control
164 ± 79 ± 54 ± 39 ± 16.6 ± 7.46 ± 110 ± 98 ± 34 ± 24 ±
13 7 13 2 1.3
0.03 11 0.5 2 2
Group 2: Cannula
156 ± 68 ± 47 ± 30 ± 13.3 ± 7.32 ± 151 ± 98 ± 38 ± 19 ±
30 30 18 5* 1.9* 0.12* 45 2 9 4
Group 3: Ventricle
130 ± 20 67 ± 13 51 ± 8 32 ± 5* 13.2 ± 1.2* 7.37 ± 0.03 190±36t 99 ± 0.3 34 ± 6 21 ± 4
Values represent mean ± standard deviation (N = 5). BP. Blood pressure; Hgb, hemoglobin; Pao-, arterial oxygen tension; Paco-, arterial carbon dioxide tension: HeO l , bicarbonate radical. of'
tf'
< 0.05 < 0.01
when compared with control. when compared with control.
Table II. Experimental variables during blood flow distribution studies Parameter
Group I: Control
Weight (kg) Anesthesia duration (min) Bypass duration (min) Flow (ml/kg/min)
3.46 ± 0.20 200 ± 55 N/A N/A
Group 2: Cannula
4.01 ± 249 ± 121 ± 29.4 ±
0.50 86 56 5.4
Group 3: Ventricle
3.90 ± 229 ± 126 ± 30.2 ±
0.82 58 51 0.2
Values represent mean ± standard deviation (N = 5). Nj A, Not applicable.
jecting microspheres into the LV via a catheter placed in the ligated common carotid artery. Also using a similar technique in piglets, Laptook, Stonestreet, and William19 demonstrated that ligation of the common carotid artery does not alter total blood flow to the brain. The effects of ligation of both the carotid artery and the jugular vein, however, have not been studied previously. Data from this study indicate that regional blood flow distribution during ECMO is not the same as when the entire cardiac output is delivered from the LV (Table III). Even the regional distribution of blood from the output of the LV is altered when blood is entering the aorta from the perfusion cannula. Distribution of 99mTc_MAA to the total brain was decreased during ECMO as compared with control when it originated in either the LV or the perfusion cannula. In the rabbit, the innominate artery branches into the left and right common carotid arteries and the right subclavian artery (Fig. 2). The hindbrain is supplied mainly by the vertebral arteries.P If flow from the ECMO circuit travels preferentially cranially via the carotids and caudally via the descending aorta, this could explain the decreased hindbrain flow in both bypass groups. The subclavian arteries may not be fed by the ECMO circuit, or perhaps the right subclavian might be supplied by the ECMO circuit while the left is not. Because of the rabbit anatomy, the possibility also exists
that blood perfusing the brain may come predominantly from the ECMO circuit while total flow is still less than control. Hemodilution has been shown to cause an increased CBF, probably related to a decreased viscosity.2l,22 An increase in CBF could compensate for the diminished oxygen-carrying capacity.F" 24 Despite these phenomena, in this experiment, in addition to a decreased relative blood flow to the brain, there resulted a decreased relative total oxygen delivery because of the significantly lowered hematocrit value. The myocardium may also have a lowered total oxygen delivery during ECMO. The decreased percentage of counts in the heart when the 99mTc-MAA originated in the ECMO circuit, as compared with when it originated in the LV in the control group, suggested that the myocardium during ECMO is not well perfused with blood from the ECMO circuit. This is supported by the fact that 99mTc_MAA distribution to the heart from the perfusion cannula during ECMO was also significantly less than that exiting the LV during ECMO. Rabbit great vessel anatomy is similar enough to the human anatomy to conclude that perfusion of the heart in the human via the low-flow ECMO circuit might also be decreased. Data from this study suggest that blood from the ECMO circuit may not pass retrogradely down the ascending aorta in sufficient amounts. This hypothesis is supported by
The Journal of Thoracic and Cardiovascular Surgery
Nowlen et al.
1 1 42
Table III. Regional blood flow distribution in controls and rabbits on ECMO (reported as percent of total counts) Organ Left brain Right brain Hindbrain Total brain Heart Left lung Right lung Left kidney Right kidney Stomach Intestine Liver Spleen Musculoskeletal (carcass)
Group 1: Control
3.37 ± 3.05 ± 3.32 ± 9.75 ± 3.95 ± 2.93 ± 3.05 ± 8.94 ± 8.68 ± 4.86 ± 10.56 ± 7.67 ± 3.49 ± 35.93 ±
1.03 1.00 1.17 3.00 1.11 0.93 0.98 1.76 1.6I 0.56 1.74 2.43 1.30 4.66
Group 2: Cannula
2.27 ± 2.20 ± 1.89 ± 6.36 ± 1.79 ± 2.08 ± 2.05 ± 6.53 ± 6.22 ± 7.81 ± 9.29 ± 7.53 ± 2.52 ± 47.83 ±
0.46* 0.44 0.25* 1.13* 0.39t:j: 0.50 0.30* 1.93 2.07 4.57 I.83§ 2.70 . 0.48 8.08*
Group 3: Ventricle
2.01 ± 0.57* 1.96 ± 0.58 1.92 ± 0.70* 5.89 ± 1.83* 3.17±0.78 1.59 ± 0.66* 1.65 ± 0.71* 7.91 ± 1.83 7.68 ± 1.83 7.12 ± 3.09 14.50 ± 2.01* 8.15 ± 1.16 1.81 ± 0.98* 40.54 ± 5.17
Values represent mean ± standard deviation (N : 5) .
• p < 0.05 when compared with control. tp
< 0.01
when compared with control.
:j:p < 0.05 when cannula is compared with ventricle.
§p
< 0.01
when cannula is compared with ventricle.
Seeker-Walker and associates.P who demonstrated during venoarterial bypass that even when 85% of the systemic output originated in the bypass circuit only 25% of the coronary artery flow originated in the bypass circuit. Of special significance during clinical ECMO is the fact that cannula blood is well oxygenated and alkalotic, whereas LV blood may be poorly oxygenated and acidotic. Thus the relative lack of cannula flow to the heart may be even more physiologically significant. An increased relative blood flow was noted to the intestine in the ventricle group as compared with the control group (Table III). Although the mechanism for the increased flow to the splanchnic regions during ECC is unknown, this phenomenon has been documented by other researchers in various animals.v 8, 21.26, 27 Distributions to the heart and intestines were the only two significant differences noted between the two bypass groups (Table III). This suggests that overall mixing of blood from the perfusion circuit and the LV (with the possible exceptions of distribution to heart and intestine) is adequate. This conclusion is supported by the observation that there were no significant differences in flow between the left and right kidneys, lungs, and cerebral hemispheres within any group. Effects of anesthesia on regional blood flow in rabbits have been reported. 28- 32 Xylazine and ketamine in the same combination used in this study were shown to produce a 30% decrease in blood pressure, a 77% decrease in respiratory rate, and a 28% reduction in heart rate." No studies were found on the regional distribution of cardiac output in the rabbit when ketamine and xylazine were used as the anesthetic. Dhasma and associa tes, 29 however,
have studied the effects of 2, 6, and 18 rug/kg doses of ketamine in rabbits. Cardiac output and blood flow to many tissues (including total brain and heart) remained unchanged. In contrast, Blake and Komer.l? using a 12 rng/kg dose of ketamine in rabbits, demonstrated an increase in arterial pressure, cardiac output, and heart rate with no significant change in total peripheral resistance. These studies suggest that the anesthetic used in this study may not affect cardiac output distribution. In the event that the anesthetic does influence regional blood flow, the effect of ECMO should still be evident, because all animals were anesthetized, and there were no significant differences in the duration of anesthesia among the three groups. The methods used in this study do not allow the determination of absolute blood flows during ECMO. Future experiments should be designed to determine absolute flows. The blood perfusion rate of 30 nil/kg/min used in this study was the maximum we could obtain with this model. This value is less than the usual flows of 100 ml/ kg/min used to fully support the human newborn infant with ECMO, so that conclusions drawn from this study must be viewed with caution. Several conclusions can be drawn from these studies. First, normal regional blood flow distribution in patients during ECMO cannot be assumed. Second, the heart and total brain may be significantly deprived of oxygenated blood during ECMO. Third, adequate mixing of blood pumped by the perfusion circuit and the heart does occur in the rabbit during low-flow ECMO with the possible exceptions of the heart and intestines. If this is indeed the case, then a better argument might be made for veno-
Volume 98 Number 6 December 1989
venous bypass in patients with lung disease alone, and special care should be given in choosing and cannulating patients with cardiac failure. REFERENCES 1. Toomasian JM, Snedecor SM, Cornell RG, Cilley RG, Bartlett RH. National experience with extracorporeal membrane oxygenation for newborn respiratory failure. ASAIO Trans 1988;34:140-7. 2. Redmond CR, Graves ED, Falterman KW, Ochsner JL, Arensman RM. Extracorporeal membrane oxygenation for respiratory and cardiac failure in infants and children. J THORAC CARDIOVASC SURG 1987;93: 199-204. 3. Kanter KR, Pennington DG, Weber TR, Zambie MA, Braun P, Martychenko V. Extracorporeal membrane oxygenation for postoperative cardiac support in children. J THoRAc CARDIOVASC SURG 1987;93:27-35. 4. Iatridis A, Voorhees M, Miller M. Review of ECMO in adults: North American experience. Perfusion 1988;3:3740. 5. Bartlett RH, Gazzaniga AB. Physiology and pathophysiology of extracorporeal circulation. In: Ionescu MI, ed. Techniques in extracorporeal circulation. 2nd ed. London: Butterworth, 1981:5-6. 6. Lees MH, Herr RH, Hill JD, et al. Distribution of systemic blood flow of the rhesus monkey during cardiopulmonary bypass. J THORAC CARDIOVASC SURG 1971;61:570-86. 7. Rudy LW, Heymann MA, Edmunds LH. Distribution of systemic blood flowduring total cardiopulmonary bypass in rhesus monkeys. Surg Forum 1970;21: 149-51. 8. Rudy LW, Heymann MA, Edmunds LH. Distribution of systemic blood flowduring cardiopulmonary bypass. J Appl Physiol 1973;34:194-200. 9. Sorenson HR, Husum B, Waaben J, et al. Brain microvascular function during cardiopulmonary bypass. J THORAC CARDIOVASC SURG 1987;94:727-32. 10. Govier A V, Reves JG, McKay RD, et al. Factors and their influence on regional cerebral blood flow during nonpulsatile cardiopulmonary bypass. Ann Thorac Surg 1984; 38:592-600. II. Santillan GG, Chemnitius JM, Bing RJ. The effect of cardiopulmonary bypass on cerebral blood flow. Brain Res 1985;345: 1-9. 12. Henriksen L, Hjems E, Lindeburgh T. Brain hyperperfusion during cardiac operations. J THORAC CARDIOVASC SURG 1983;86:202-8. 1'3. Lundar T, Lindegaard KF, Froysaker T, Aaslid R, Grip A, Nornes H. Dissociation between cerebral autoregulation and carbon dioxide reactivity during nonpulsatile cardiopulmonary bypass. Ann Thorac Surg 1985;40:582-7. 14. Lundar T, Lindegaard KF, Froysaker T, Aaslid R, Wiberg J, Nornes H. Cerebral perfusion during nonpulsatile cardiopulmonary bypass. Ann Thorac Surg 1985;40: 144-50. 15. Patterson JL. Circulation through the brain. In: Ruch TC, Patton HD, eds. Physiology and biophysics. Philadelphia: Saunders, 1965:953-8. 16. Furzan J, Gabriele G, Wheeler J, Fixler D, Rosenfeld e.
Regional blood flow distribution during EeMO
I 1 43
Regional blood flows in newborn lambs during endotracheal continuous airway pressure and continuous negative pressure breathing. Pediatr Res 1981;15:874-8. 17. Mendell P, Hollenberg N. Cardiac output distribution in the rat: comparison of rubidium and microsphere methods. Am J PhysioI1971;221:1617-20. 18. Neutze JM, Wyler F, Rulolph AM. Use of radioactive rnicrospheres to assess distribution of cardiac output in rabbits. Am J Physiol 1968;215:486-95. 19. Laptook AR, Stonestreet BS, William OH. The effect of carotid artery ligation on brain blood flow in newborn piglets. Brain Res 1983;276:51-4. 20. Baldwin BA, Bell FR. The anatomy of the cerebral circulation of the sheep and ox: the dynamic distribution of the blood supplied by the carotid and vertebral arteries to the cranial regions. J Anat 1963;97:203-15. 21. Utley JR, Wachtel C, Cain RB, Spaw EA, Collins JC, Stephens DB. Effects of hypothermia, hemodilutin, and pump oxygenation on organ water content, blood flow and oxygen delivery, and renal function. Ann Thorac Surg 1981;31:121-33. 22. Gordon RJ, Ravin M, DaicoffGR, Rawitscher RE. Effects of hemodilution on hypotension during cardiopulmonary bypass. Anesth Analg 1975;54:482-8. 23. Tanaka T, Inoue T, Paton Be. Oxygen availability during hypothermic perfusion using diluted blood. Surg Forum 1962;13:138-40. 24. Wright CJ. The effects of severe progressive hemodilution on regional blood flow and oxygen consumption. Surgery 1976;79:299-305. 25. Seeker-Walker JS, Edmonds JF, Spratt EH, Conn A W. The source of coronary perfusion during partial bypass for extracorporeal membrane oxygenation (ECMO). Ann Thorac Surg 1976;21: 138-43. 26. Rudy LW, Boucher JK, Edmunds LH Jr. The effect of deep hypothermia and circulatory arrest on the distribution of systemic blood flow in rhesus monkeys. J THORAC CARDIOVASC SURG 1972;64:706-13. 27. Halley MM, Reemtsma K, Creech 0 Jr. Hemodynamics and metabolism of individual organs during extracorporeal circulation. Surgery 1959;46: 1128-34. 28. Sanford TD, Colby ED. Effect of xylazine and ketamine on blood pressure, heart rate and respiratory rate in rabbits. Lab Anim Sci 1980;30:519-23. 29. Dhasma KM, Saxena PR, Prakash 0, Van Der Zee HT. A study on the influence of ketamine on systemic and regional haemodynamics in conscious rabbits. Arch Int Pharmacodyn 1984;269:323-34. 30. Blake DW, Korner PI. Effects of ketamine and althesin anesthesia on baroreceptor-heart rate reflex and hemodynamics of intact and pontine rabbits. J Auton Nerv Syst 1982;5:145-54. 31. Wyler F, Neutze JM, Rudolph AM. Effects of endotoxin on distribution of cardiac output in unanesthetized rabbits. Am J Physiol 1970;219:246-51. 32. Wyler F. Effect of general anesthesia on distribution of cardiac output and organ blood flow in the rabbit: halothane and chloralose-urethane. J Surg Res 1974;17:381-6.