- Email: [email protected]
Arteriovenous ECMO for neonatal respiratory support
Arteriovenous ECMO for neonatal respiratory support A study in perigestational lambs A study was undertaken to investigate the applicability of the arteriovenous mode of perfusion for partial support of neonatal respiration. Perigestational lambs, delivered by cesarean section, served as the animal model of respiratory distress. Arteriovenous flow was accomplished between a single umbilical artery and vein. A microchannel membrane oxygenator was used to provide partial respiratory support to the newborn lambs. Total systemic flow, pulmonary blood flow, and pulmonary vascular resistance were assessed at various rates of arteriovenous perfusion and correlated with systemic oxygenation. A reduction in right-to-left shunting of blood and pulmonary vascular resistance occurred as arterial oxygenation rose from conditions of hypoxemia to Pan, values higher than 50 torr. Myocardial performance was not impaired at rates of arteriovenous perfusion below 30 percent of the total systemic fiow, as evidenced by normal electrocardiographic tracings, pulmonary capillary wedge pressures, and central venous pressures. Arteriovenous extracorporeal membrane oxgenation (ECMO) may be particularly suitable for use in infants with hypoxia and high pulmonary vascular resistance.
Bartley P. Griffith, M.D.,* Harvey S. Borovetz, Ph.D.,** Robert L. Hardesty, M.D.,* Tin-Kan Hung, Ph.D.,*** and Henry T. Bahnson, M.D.,* Pittsburgh, Pa.
Extracorporeal membrane oxygenation (ECMa) recently has been employed to support infants with respiratory distress. i. 2 Conventionally, the venoarterial mode of extracorporeal perfusion is used, in which venous blood is drained from the right atrium, oxygenated, and returned to a central artery. Although this technique provides both circulatory and pulmonary support, arteriovenous ECMa is an attractive alternative when used in combination with positive-pressure respiration for partial pulmonary support in neonates. From the Surgical Research Laboratory, University of Pittsburgh School of Medicine, Pittsburgh, Pa. Supported in part by grant-in-aid awards from the Western Pennsylvania Heart Association and Health Research and Services Foundation. Bartley P. Griffith, M.D., is the recipient of a Research Fellowship Award from the American Heart Association. Received for publication Aug. 17, 1978. Accepted for publication Nov. 15, 1978. Address for reprints: Dr. Bartley P. Griffith University of Pittsburgh School of Medicine, Department of Surgery, Pittsburgh. Pa. 15261. "Department of Surgery. --Departments of Surgery and Civil Engineering. ·-·Departments of Civil Engineering and Neurological Surgery.
Arteriovenous bypass mimics placental circulation and may reduce pulmonary vascular resistance by delivery of oxygenated blood directly to the precapillary sphincters of the pulmonary arterioles ."? Umbilical vessels may be used for vascular access, so that the jugular vein and central artery (carotid or axillary) need not be sacrificed." In an extracorporeal circuit with low resistance to flow, arteriovenous perfusion may be powered by the heart rather than by a mechanical pump. A study was undertaken to investigate the applicability of the arteriovenous mode of perfusion with ECMa to partial support of neonatal respiration. Perigestational lambs, delivered by cesarean section, were used for the study. Lambs were chosen since their weight at birth approximates that of a newborn infant and the incomplete gestation ensures respiratory distress. Methods A cesarean section was performed on 10 pregnant ewes to deliver premature lambs (0.8 to 0.9 of term) weighing 2.5 to 3.5 kilograms. Regional anesthesia (0.5 percent lidocaine) was administered. The fetal pelvis was palpated through the uterus and served as the
0022-5223/79/040595+07$00.70/0 © 1979 The C. V. Mosby Co.
595
The Journal of
596
Griffith et al.
Thoracic and Cardiovascular Surgery
Fig. 1. Radiograph showing placement of umbilical cannulas and pressure-monitoring lines in the neonatal lamb.
landmark for the hysterotomy. Once delivered, the lamb was placed on the abdomen of the ewe, without obstructing or stretching the umbilical vessels, and was immediately intubated with a No. 18 Fr. endotracheal tube. Irregular spontaneous respirations occurred upon delivery. Catheters were placed in the aortic arch and right atrium through a femoral artery and vein, and pressures were measured on Statham transducers (P-23 series). An umbilical artery and vein were isolated. A polyurethane catheter (outer diameter 8 Fr., inner diameter 6 Fr.) reinforced with thin wire? and fitted with an inner stylet was introduced through an arteriotomy after the artery was physically dilated . The catheter was coated with a surgical lubricant to compensate for vascular spasm and friability. The stylet was removed as the arterial catheter was advanced beyond the acute angle formed where the umbilical artery passed retroperitoneally. The arterial catheter is shown proximal to the bifurcation of the aorta in Fig . I. A similar technique was performed on the umbilical vein, and the catheter was positioned 3 em. into this vessel (Fig . I) . A period of 10 to 20 minutes was required for cannulation, during which time partial placental circulation through the remaining umbilical artery and vein supported inadequate spontaneous respirations. The extracorporeal circuit consisted of a microchannel oxygenator with 0.6 sq. M. of membrane area for gas exchange, a small roller pump,* a bubble trap, and interconnecting Tygon tubing. The microchannel oxygenator was designed to provide partial
support of respiration to infants and small children who cannot be oxygenated or ventilated adequately by positive-pressure ventilation alone ." Among the desirable characteristics of the microchannel design are (l) efficient oxygen and carbon dioxide exchange," (2) a low priming volume (80 ml. for a unit capable of oxygenating 1.4 L. of blood per minute), (3) a potential for applying TDMAC* or vacuum vapor deposited carbont to surfaces in contact with blood in order to achieve thromboresistance , 10 and (4) a low resistance to the flow of blood, which makes flow possible without an extracorporeal pump. A roller pump was employed in the present study solely to guarantee a constant rate of arteriovenous flow for acquisition of data. All interconnecting tubing was shortened to minimize the priming volume (150 to 220 c.c.) . The priming solution consisted of whole blood which was anticoagulated with heparin and balanced appropriately for pH, Na", K+, and Ca"" . Before the start of arteriovenous ECMO, the remaining umbilical vessels were ligated, the cord was severed, and the animal was positioned adjacent to the ECMO circuit (Fig. 2) . The lamb was paralyzed with pancuronium (0.1 mg . per kilogram) and ventilated on room air with a tidal volume of 30 ml. per kilogram and at a respiratory rate between 18 and 30, chosen to maintain normal Paco.. Transfusions of whole blood were given to maintain the peak arterial pressure above 50 torr and the central venous pressure between 5 and 10
*Cole-Parmer Co., Chicago, Ill.
tBiolite, Carbo Medics, Inc., San Diego, Calif .
*Tridodeclymethylammoniurn chloride, Polysciences Corp., Warrington , Pa.
Volume
rr
Arteriovenous ECMO
Number 4 April,1979
597
CD ~:::2:::;"lI
~ --l'f~~=~~~~
TO OXYGENATOR INLET MANIFOLD
FROM OXYGENATOR OUTLET MANIFOLD
c:
:@ ~
1.
Fig. 2. Perfusion circuit for arteriovenous extracorporeal membrane oxygenation. Oxygen was introduced into the inlet manifold of the microchannel oxygenator at a rate of 5 L. per minute. Nomenclature: I, Bourn's respirator. 2, Cardiac output computer. 3, Roller pump. 4, Densitometer and minicircuit. 5, Microchannel membraneoxygenator. 6. Bubble trap.
torr. Sodium bicarbonate was administered to correct acidosis. A No.5 Fr. balloon-tip catheter (Swan-Ganz) was inserted via the right external jugular vein into the pulmonary artery. The position of the catheter was verified by pulse contours and later by examination at autopsy. Heparin was administered systemically at an initial concentration of 5 mg. per kilogram to maintain the whole blood activated clotting time at three to five times the control value. Rectal temperature was maintained at 36° ± 3° C. by heating blankets and infrared lamps. The electrocardiogram was monitored with the standard limb leads. Arteriovenous flow then was initiated for periods of 4 to 6 hours (Fig. 2). Indocyanine green dye curves were used qualitatively to demonstrate patterns of blood flow and to estimate total systemic flow. The latter is the sum of blood flows through the patent foramen ovale, patent ductus arteriosus, and the blood which is passed normally through the circulation. A continuous sampling method was designed specifically to avoid loss of volume in small animals. 11 A circuit including a densitometer and small roller pump was positioned in parallel with the inlet tubing to the oxygenator (Fig. 2). Blood was delivered through this minicircuit at a constant rate of 20 ml per minute. Next, 1.25 mg. of dye was injected into the right atrium, and concentration changes at the distal aorta were measured by the densitometer in the minicircuit. A dye curve with two peaks was inscribed. The initial peak represented that portion
20
40
Fig. 3. Oxygenationand hemodynamicparameters monitored during a 140 minute period of a 5 hour ECMO study. Arteriovenous ECMO flow (Qav J was increased from zero to rates equalling 20 to 30 percent of the total systemic flow (QtJ, i.e., Qav/Qt = 0.2 to 0.3. Arteriovenous flow was stopped periodically to check base-line measurements.
of the circulation that was shunted right to left; the second peak was produced by the normal circulation of blood. The forward triangle method was used to calculate total systemic flow (or venous return) from the area inscribed by both peaks. 12 The right-to-left shunting of blood through the patent foramen ovale and ductus arteriosus was expressed as a percentage of the total systemic flow. 12 Pulmonary blood flow (PBF) was determined from the difference between total systemic flow and the right-to-left shunt. Pulmonary vascular resistance (PVR)13 was calculated from the formula PVR
=
PAP - PCWP PBF
where PAP is mean pulmonary artery pressure and PCWP is mean pulmonary capillary wedge pressure. Rates of flow through the oxygenator (Qav) were determined by precalibrating the flow generated by the ECMO pump. Data obtained for right-to-left shunting and PVR were correlated with the oxygen tension in the aorta
598
The Journalof Thoracic and Cardiovascular Surgery
Griffith et al.
Table I. Arterial oxygenation as measured for the perfusion experiment shown in Fig. 3 Pulmonary artery" QavlQt (%)
pH
0 7 11 15 17 20 0 9
7.40 7.48 7.52 7.55 7.55 7.58 7.34 7.40 7.47 7.50 7.54 7.36 7.57
11
24 34 0 44
I
Po, (torr)
29 52 57 62 60 57 17 43 50 52 57 18 60
I
Aorta" Pea, (torr)
pH
35 30 28 26 25 22 42 34 29 25 19 45 21
7.50 7.51 7.49 7.54 7.58 7.57 7.36 7.38 7.46 7.50 7.53 7.42 7.49
I
Po, (torr) 44
59 72
78 75 72
31 50 60 64
70 27 64
I
Pea, (torr)
29 34 30 25 24 21 40 36 30 25 18 40 23
Oxygen transfers (Cc. Imin.)
1.4 1.4 1.7 2.2 2.2 2.1 2.5 3.0 2.5 4.5
Legend: Qav/Qt, Ratio of flow through the oxygenator (Qav) to total systemic flow (Qt). ·Measured at 37" C. tCalculated for the microchannel oxygenator."
(Pao,,). Blood gases were measured with Radiometer microequipment (Type E 5021). The t test for paired observations with equal variance was used to calculate the significance of changes in the percentage of flow shunted right to left and PVR at various levels of aortic oxygenation.
Results Experimental data were obtained from eight of the 10 newborn lambs. Technical problems early in the development of the animal model associated with the cannulation of the umbilical vessels resulted in prolonged acidosis and death of two of the lambs before arteriovenous support could be initiated. The delivery of arteriovenous flow to the ECMO circuit requires satisfactory myocardial performance. The limiting factors for establishing satisfactory arteriovenous ECMO flows were mean arterial pressure in excess of 30 torr and central venous pressure between 4 and 10 torr. A linear relationship existed between changes in arteriovenous ECMO and total systemic flow rates. The increase in total systemic flow was a reflection of an increase in venous return commensurate with the arteriovenous ECMO rate of perfusion. Similarly, pulmonary and systemic arterial P0 2 varied directly with the arteriovenous rate of perfusion. At the higher arterial P0 2 values (> 50 torr), PBF rose while PVR and the percentage of flow shunted right to left decreased. Arteriovenous ECMO was begun at rates of flow of 20 to 30 ml per kilogram per minute. At these low rates, an initial 7 to 15 torr reduction in peak arterial
pressure from 50 to 65 torr was noted in all experiments. Simultaneously, the central venous pressure fell below 5 torr. Transfusions of 50 to 150 ml. of whole blood raised the central venous pressure to 5 to 10 torr and produced stable peak arterial pressures above 50 torr. Arteriovenous ECMO was continued at these rates of flow for 15 to 30 minutes in order to permit cardiovascular accommodation. Perfusion was subsequently increased incrementally to values approaching 100 to 150 c.c. per kilogram per minute. The variations in aortic and pulmonary arterial P0 2 • total systemic flow (Qt), PBF, and PVR at different rates of arteriovenous flow are shown for a typical experiment in Fig. 3. In the absence of arteriovenous perfusion (Qav = 0), arterial oxygenation was severely depressed and PVR was three times the normal value for lambs at term. As arteriovenous perfusion was increased from zero conditions to rates equalling 20 to 30 percent of the total systemic flow (Qav / Qt = 0.2 to 0.3), pulmonary artery P0 2 rose from base-line values of 17 to 35 torr to 52 to 60 torr. A corresponding rise in P0 2 was measured in the aorta (Table I). The elimination of carbon dioxide was proportional to the arteriovenous ECMO flow: At Qav/ Qt > 0.2, Paco" was reduced by 15 to 20 torr from the base-line conditions in which Qav = 0 (Table I). Right-to-left shunting of blood through the patent foramen ovale and ductus arteriosus varied with arterial oxygenation (Fig. 4). The data compiled from the eight animals of this study were taken for Qav/Qt ratios ranging from zero to 45 percent. A reduction in the percentage of flow shunted right to left under conditions of
Volume 77
Arteriovenous ECMO
Number 4
599
April, 1979
50
I!
p
2,5r-----.-----.-----.-----,
t p<~7' p
10
p
30-38
40-49
50-64
70-112
Pc,02-Torr
Fig. 4. Variation in the right-to-left shuntingof bloodthrough the patent foramen ovale and ductus arteriosus with arterial oxygenation. Results are plotted as the mean ± S.E.M. for data compiled from all animals of this study. hypoxemia (Pao, = 30 to 38 torr) occurred at higher levels of Pao2 (>50 torr) i.e., from 38 percent to 20 to 28 percent. This reduction is significant (p < 0.00 1). An inverse relationship existed between the PVR and level of systemic arterial oxygenation (Fig. 5). PVR was 40 percent lower at Pao2 > 66 torr than at Pao. < 38 torr. The average pulmonary artery pressure was not elevated from prebypass conditions (48/27 torr) at arteriovenous ECMO flows corresponding to less than 30 percent of the total systemic flow (p> 0.05). Discussion Arteriovenous ECMO has been used by others for isolated fetal perfusion. 14-16 Rashkind and associates'? employed arteriovenous perfusion with a bubble oxygenator for gas exchange to support four neonates with hyaline membrane disease and four children with respiratory failure due to cystic fibrosis. Dorson and colleagues" demonstrated the applicability of cannulating the umbilical vessels for arteriovenous ECMO during perfusion of five neonates in respiratory distress. In the present study, experiments were conducted to identify the acute hemodynamic response to arteriovenous perfusion. The results of these acute experiments can be used as a guide for an improved "chronic" mode!. A single distal aortic annul a inserted from an umbilical artery sustained blood flows in excess of 100 m!. per killogram per minute for periods of up to 6 hours. The bypass rates of flow were periodically between 175 and 200 m!. per kilogram per minute. The pressure drop across the polyurethane catheter at flow rates equal to
0-5'----..L..---....I..------'-----' 31-38 40-49 55-64 66-112
PcP2 -Torr
Fig. 5. Reduction in pulmonary vascular resistance with increasinglevelsof arterial oxygenation. The data are plottedas described in Fig. 4. 100 m!. per kilogram per minute was estimated from Poiseuille's law to be 30 torr. Thus the arterial pressure required to establish arteriovenous perfusion at these flows is 30 torr. In the absence of a mechanical pump, this value would be increased to 40 to 55 torr in order to include the low resistance to flow in the microchannel oxygenator (10 to 25 torr). The rate of oxygen transferred by the microchannel oxygenator at arteriovenous ECMO flows between 20 and 45 percent of the total systemic flow ranged between 1.4 and 4.5 C.c. of oxygen per minute (Table I). This figure compares with the oxygen requirement for lambs at 130 days of gestation of 18 c.c. per minute for normothermic conditions. 14 For lambs in the present study which were paralyzed and hypothermic the rate of oxygen delivery in the extracorporeal circuit was sufficient to increase arterial oxygenation to Iifesustaining levels. The therapeutic effects of delivering highly oxygenated blood to the pulmonary vasculature include a reduction in the percentage of flow shunted right to left, an elevation in PBF, and a fall in PVR (Figs. 4 and 5). The fall in PVR with an elevation of PBF suggests that there occurs a vasodilatory effect at the precapillary sphincter mediated by elevation in pulmonary arterial oxygenation." When pulmonary and systemic arterial pressures are the same, the ratio of flow through the ductus to that in the pulmonary circulation varies directly with the systemic vascular resistance and PVR. In neonates with elevated pulmonary vascular resistance and equal pulmonary and systemic arterial pressures, a fetal pattern of circulation exists, characterized by marked right-to-Ieft shunting of blood
The Journal of
6 0 0 Griffith et al.
through the ductus, diminished PBF and consequential systemic hypoxia. The partial respiratory support provided by arteriovenous ECMO causes a therapeutic reversal in this circulatory pattern through delivery of sufficient increments of oxygen to the pulmonary vasculature. Zapol and Kolobow!" presented cineangiographic evidence that the diameter of the ductus arteriosus narrowed markedly at arterial P0 2 values above 35 torr. This implies that the reduction in right-to-left shunting observed in the present study at Pao. above 50 torr may be in part a consequence of the narrowing of the ductus. The reversal in fetal pattern rapidly occurred following the initiation of extracorporeal support and was dependent upon the continuation of ECMO (Fig. 3). The use of arteriovenous ECMO has been restricted due to theories that an increase in venous return leads to an increase in cardiac output, which is deleterious to the myocardium and especially to the right ventricle should pulmonary hypertension exist. 2 However, there is evidence to suggest that partial respiratory support with arteriovenous ECMO can be tolerated by the myocardium of a newborn infant. Rudolph 18 has described the primary change in the circulation of newborn infants at birth as a shift in the flow of blood for gas exchange from the placenta to the lungs. He found right ventricular output in utero to be 300 ml. per kilogram per minute, with 200 rnl. of the flow through the descending aorta being delivered to the placenta. Separation of the infant from the umbilical-placental circulation results in a rapid decrease in venous return and right ventricular output by 200 m!. per kilogram per minute. Postnatally satisfactory oxygenation can be provided by arteriovenous ECMO at rates of flow (Qav < 100 ml, per kilogram per minute) that do not exceed those normally diverted to the placenta. Consequently, right ventricular output should not exceed the flow which previously existed in utero. The results of the present study and studies reported previously": 20 suggest that an initial rise in cardiac output occurs during arteriovenous perfusion in direct response to an increase in the venous return (Fig. 3). This rise appears to be transitory and the cardiac output returns to prebypass levels within I to 2 hours after initiation of arteriovenous ECMO at a constant rate (Qav/Qt :5 0.3). Pulmonary arterial pressure is not elevated and PVR is reduced at those flows which increase arterial oxygenation to life-sustaining levels (Pao. > 50 torr). The lack of variation in electrocardiogram, pulmonary capillary wedge, and central venous pressures indicates that myocardial performance is not impaired.
Thoracic and Cardiovascular Surgery
Partial respiratory support provided by arteriovenous ECMO with the microchannel oxygenator appears to offer special therapeutic advantages in neonates with severe respiratory disease resulting in a fetal circulatory pattern. We wish to acknowledge the design concept for the microchannel oxygenator by Michael H. Weissman, M.D.. Ph.D., in 1969 at Carnegie-Mellon University and to thank Frank McSteen, B.S., for assistance in conducting the animal experiments. REFERENCES Gazzaniga AB, Fong SW, Jefferies MR, Roohk HV, Haiduc M, Bartlett RH: Extracorporeal membrane oxygenator support for cardiopulmonary failure. J THORAC CARDIOVASC SURG 73:375-386, 1976 2 Bartlett RH, Gazzaniga AB: Extracorporeal circulation for cardiopulmonary failure. Curr Probl Surg 15: 1-96. 1978 3 Bohr OF: The pulmonary hypoxic response. Chest 71:244-246, 1977 4 Fishman AP: Hypoxia on the pulmonary circulation. How and where it acts. Circ Res 38: 221-231, 1976 5 Detar R, Gella M: Oxygen and isolated vascular smooth muscle from the main pulmonary artery of the rabbit. Am J Physiol 221:1791-1794, 1971 6 Dorson W, Heyer B, Baker E, Cohen M, Elgis R, Molthan M, Trump 0: Response of distressed infants to partial bypass lung assist. Trans Am Soc Artif Intern Organs 16:345-361, 1970 7 Kolobow T, Zapol W: A new thin-walled non-kinking catheter for peripheral vascular cannulation. Surgery 68:625-629, 1970 8 Hung T-K, Borovetz HS, Hardesty RL, Weissman MH: Toward the development of a neonatal pulmonary assist membrane oxygenator. S Afr Mech Engin 28: 112-116, 1978 9 Hung T-K, Borovetz HS, Weissman MH: Transport and flow phenomena in a microchannel membrane oxygenator. Ann Biomed Engin 5:343-361, 1977 10. Borovetz HS, Griffith BP, Phillips LV, Haubold AD, Hercules, OM, Hung T-K, Hardesty RL: Scanning electron microscopic and surface analytic study of an isotropic vacuum vapor deposited carbon on microporous membranes. Scanning Electron Microosc 11:85-94, 1978 II Griffith BP, Borovetz HS, Hung T-K, Hardesty RL: Continuous indocyanine green dye determination of cardiac output through an arteriovenous shunt. Proc Assoc Advancement Med Inst, Thirteenth Annual Meeting, p 109, 1978 12 Rudolph AM: Cardiac catheterization and angiocardiography, Congenital Diseases of the Heart, Chicago, 1974, Year Book Medical Publishers, Inc., pp 139-149 13 Guyton AC: Textbook of Medical Physiology, Philadelphia, 1971, W. B. Saunders Company, p 206
Volume
rt
Arteriovenous ECMO
Number 4
60 1
April, 1979
14 Zapol W, Kolobow T: Isolated extracorporeal placentation of the fetal lamb, Mammalian Fetus in vitro, CR Austin, ed., London, 1973, Chapman & Hall, Ltd., pp 147-193 15 Doppman JL, Zapol W, Kolobow T, Pierce J: Angiocardiography of fetal lambs on artificial placentas. Invest Radiol 5:181-186, 1970 16 Alexander DP, Britton HG, Nixon DA: Survival of the foetal sheep at term following short periods of perfusion through the umbilical vessels. J Physiol (London) 175:113-124, 1964 17 Rashkind WJ, Milles WW, Falcone D, Taft RW: Hemodynamic effects of arteriovenous oxygenation with a
small volume artificial extracorporeal lung. J Pediatr 70:425-429, 1967 18 Rudolph AM: Changes in the circulation after birth, Congenital Diseases of the Heart, New York, 1974, Year Book Medical Publishers, Inc., pp 17-28 19 Awad JA, Morin PJ: Arteriovenous partial perfusion, Artificial Lungs for Acute Respiratory Failure, WM Zapol, J Qvist, eds., New York, 1976, Academic Press, Inc., pp 329-342 20 Borovetz HS, Griffith BP, Hung T-K, Hardesty RL, Weissman MH: Arteriovenous perfusion with the pulmonary assist membrane oxygenator. Int J Artif Organs 1:232-238, 1978
Information for authors Most of the provisions of the Copyright Act of 1976 became effective on January I, 1978. Therefore, all manuscripts must be accompanied by the following written statement, signed by one author: "The undersigned author transfers all copyright ownership of the manuscript (title of article) to The C. V. Mosby Company in the event the work is published. The undersigned author warrants that the article is original, is not under consideration by another journal, and has not been previously published. I sign for and accept responsibility for releasing this material on behalf of any and all co-authors." Authors will be consulted, when possible, regarding republication of their material.
Bartley P. Griffith, M.D.,* Harvey S. Borovetz, Ph.D.,** Robert L. Hardesty, M.D.,* Tin-Kan Hung, Ph.D.,*** and Henry T. Bahnson, M.D.,* Pittsburgh, Pa.
Extracorporeal membrane oxygenation (ECMa) recently has been employed to support infants with respiratory distress. i. 2 Conventionally, the venoarterial mode of extracorporeal perfusion is used, in which venous blood is drained from the right atrium, oxygenated, and returned to a central artery. Although this technique provides both circulatory and pulmonary support, arteriovenous ECMa is an attractive alternative when used in combination with positive-pressure respiration for partial pulmonary support in neonates. From the Surgical Research Laboratory, University of Pittsburgh School of Medicine, Pittsburgh, Pa. Supported in part by grant-in-aid awards from the Western Pennsylvania Heart Association and Health Research and Services Foundation. Bartley P. Griffith, M.D., is the recipient of a Research Fellowship Award from the American Heart Association. Received for publication Aug. 17, 1978. Accepted for publication Nov. 15, 1978. Address for reprints: Dr. Bartley P. Griffith University of Pittsburgh School of Medicine, Department of Surgery, Pittsburgh. Pa. 15261. "Department of Surgery. --Departments of Surgery and Civil Engineering. ·-·Departments of Civil Engineering and Neurological Surgery.
Arteriovenous bypass mimics placental circulation and may reduce pulmonary vascular resistance by delivery of oxygenated blood directly to the precapillary sphincters of the pulmonary arterioles ."? Umbilical vessels may be used for vascular access, so that the jugular vein and central artery (carotid or axillary) need not be sacrificed." In an extracorporeal circuit with low resistance to flow, arteriovenous perfusion may be powered by the heart rather than by a mechanical pump. A study was undertaken to investigate the applicability of the arteriovenous mode of perfusion with ECMa to partial support of neonatal respiration. Perigestational lambs, delivered by cesarean section, were used for the study. Lambs were chosen since their weight at birth approximates that of a newborn infant and the incomplete gestation ensures respiratory distress. Methods A cesarean section was performed on 10 pregnant ewes to deliver premature lambs (0.8 to 0.9 of term) weighing 2.5 to 3.5 kilograms. Regional anesthesia (0.5 percent lidocaine) was administered. The fetal pelvis was palpated through the uterus and served as the
0022-5223/79/040595+07$00.70/0 © 1979 The C. V. Mosby Co.
595
The Journal of
596
Griffith et al.
Thoracic and Cardiovascular Surgery
Fig. 1. Radiograph showing placement of umbilical cannulas and pressure-monitoring lines in the neonatal lamb.
landmark for the hysterotomy. Once delivered, the lamb was placed on the abdomen of the ewe, without obstructing or stretching the umbilical vessels, and was immediately intubated with a No. 18 Fr. endotracheal tube. Irregular spontaneous respirations occurred upon delivery. Catheters were placed in the aortic arch and right atrium through a femoral artery and vein, and pressures were measured on Statham transducers (P-23 series). An umbilical artery and vein were isolated. A polyurethane catheter (outer diameter 8 Fr., inner diameter 6 Fr.) reinforced with thin wire? and fitted with an inner stylet was introduced through an arteriotomy after the artery was physically dilated . The catheter was coated with a surgical lubricant to compensate for vascular spasm and friability. The stylet was removed as the arterial catheter was advanced beyond the acute angle formed where the umbilical artery passed retroperitoneally. The arterial catheter is shown proximal to the bifurcation of the aorta in Fig . I. A similar technique was performed on the umbilical vein, and the catheter was positioned 3 em. into this vessel (Fig . I) . A period of 10 to 20 minutes was required for cannulation, during which time partial placental circulation through the remaining umbilical artery and vein supported inadequate spontaneous respirations. The extracorporeal circuit consisted of a microchannel oxygenator with 0.6 sq. M. of membrane area for gas exchange, a small roller pump,* a bubble trap, and interconnecting Tygon tubing. The microchannel oxygenator was designed to provide partial
support of respiration to infants and small children who cannot be oxygenated or ventilated adequately by positive-pressure ventilation alone ." Among the desirable characteristics of the microchannel design are (l) efficient oxygen and carbon dioxide exchange," (2) a low priming volume (80 ml. for a unit capable of oxygenating 1.4 L. of blood per minute), (3) a potential for applying TDMAC* or vacuum vapor deposited carbont to surfaces in contact with blood in order to achieve thromboresistance , 10 and (4) a low resistance to the flow of blood, which makes flow possible without an extracorporeal pump. A roller pump was employed in the present study solely to guarantee a constant rate of arteriovenous flow for acquisition of data. All interconnecting tubing was shortened to minimize the priming volume (150 to 220 c.c.) . The priming solution consisted of whole blood which was anticoagulated with heparin and balanced appropriately for pH, Na", K+, and Ca"" . Before the start of arteriovenous ECMO, the remaining umbilical vessels were ligated, the cord was severed, and the animal was positioned adjacent to the ECMO circuit (Fig. 2) . The lamb was paralyzed with pancuronium (0.1 mg . per kilogram) and ventilated on room air with a tidal volume of 30 ml. per kilogram and at a respiratory rate between 18 and 30, chosen to maintain normal Paco.. Transfusions of whole blood were given to maintain the peak arterial pressure above 50 torr and the central venous pressure between 5 and 10
*Cole-Parmer Co., Chicago, Ill.
tBiolite, Carbo Medics, Inc., San Diego, Calif .
*Tridodeclymethylammoniurn chloride, Polysciences Corp., Warrington , Pa.
Volume
rr
Arteriovenous ECMO
Number 4 April,1979
597
CD ~:::2:::;"lI
~ --l'f~~=~~~~
TO OXYGENATOR INLET MANIFOLD
FROM OXYGENATOR OUTLET MANIFOLD
c:
:@ ~
1.
Fig. 2. Perfusion circuit for arteriovenous extracorporeal membrane oxygenation. Oxygen was introduced into the inlet manifold of the microchannel oxygenator at a rate of 5 L. per minute. Nomenclature: I, Bourn's respirator. 2, Cardiac output computer. 3, Roller pump. 4, Densitometer and minicircuit. 5, Microchannel membraneoxygenator. 6. Bubble trap.
torr. Sodium bicarbonate was administered to correct acidosis. A No.5 Fr. balloon-tip catheter (Swan-Ganz) was inserted via the right external jugular vein into the pulmonary artery. The position of the catheter was verified by pulse contours and later by examination at autopsy. Heparin was administered systemically at an initial concentration of 5 mg. per kilogram to maintain the whole blood activated clotting time at three to five times the control value. Rectal temperature was maintained at 36° ± 3° C. by heating blankets and infrared lamps. The electrocardiogram was monitored with the standard limb leads. Arteriovenous flow then was initiated for periods of 4 to 6 hours (Fig. 2). Indocyanine green dye curves were used qualitatively to demonstrate patterns of blood flow and to estimate total systemic flow. The latter is the sum of blood flows through the patent foramen ovale, patent ductus arteriosus, and the blood which is passed normally through the circulation. A continuous sampling method was designed specifically to avoid loss of volume in small animals. 11 A circuit including a densitometer and small roller pump was positioned in parallel with the inlet tubing to the oxygenator (Fig. 2). Blood was delivered through this minicircuit at a constant rate of 20 ml per minute. Next, 1.25 mg. of dye was injected into the right atrium, and concentration changes at the distal aorta were measured by the densitometer in the minicircuit. A dye curve with two peaks was inscribed. The initial peak represented that portion
20
40
Fig. 3. Oxygenationand hemodynamicparameters monitored during a 140 minute period of a 5 hour ECMO study. Arteriovenous ECMO flow (Qav J was increased from zero to rates equalling 20 to 30 percent of the total systemic flow (QtJ, i.e., Qav/Qt = 0.2 to 0.3. Arteriovenous flow was stopped periodically to check base-line measurements.
of the circulation that was shunted right to left; the second peak was produced by the normal circulation of blood. The forward triangle method was used to calculate total systemic flow (or venous return) from the area inscribed by both peaks. 12 The right-to-left shunting of blood through the patent foramen ovale and ductus arteriosus was expressed as a percentage of the total systemic flow. 12 Pulmonary blood flow (PBF) was determined from the difference between total systemic flow and the right-to-left shunt. Pulmonary vascular resistance (PVR)13 was calculated from the formula PVR
=
PAP - PCWP PBF
where PAP is mean pulmonary artery pressure and PCWP is mean pulmonary capillary wedge pressure. Rates of flow through the oxygenator (Qav) were determined by precalibrating the flow generated by the ECMO pump. Data obtained for right-to-left shunting and PVR were correlated with the oxygen tension in the aorta
598
The Journalof Thoracic and Cardiovascular Surgery
Griffith et al.
Table I. Arterial oxygenation as measured for the perfusion experiment shown in Fig. 3 Pulmonary artery" QavlQt (%)
pH
0 7 11 15 17 20 0 9
7.40 7.48 7.52 7.55 7.55 7.58 7.34 7.40 7.47 7.50 7.54 7.36 7.57
11
24 34 0 44
I
Po, (torr)
29 52 57 62 60 57 17 43 50 52 57 18 60
I
Aorta" Pea, (torr)
pH
35 30 28 26 25 22 42 34 29 25 19 45 21
7.50 7.51 7.49 7.54 7.58 7.57 7.36 7.38 7.46 7.50 7.53 7.42 7.49
I
Po, (torr) 44
59 72
78 75 72
31 50 60 64
70 27 64
I
Pea, (torr)
29 34 30 25 24 21 40 36 30 25 18 40 23
Oxygen transfers (Cc. Imin.)
1.4 1.4 1.7 2.2 2.2 2.1 2.5 3.0 2.5 4.5
Legend: Qav/Qt, Ratio of flow through the oxygenator (Qav) to total systemic flow (Qt). ·Measured at 37" C. tCalculated for the microchannel oxygenator."
(Pao,,). Blood gases were measured with Radiometer microequipment (Type E 5021). The t test for paired observations with equal variance was used to calculate the significance of changes in the percentage of flow shunted right to left and PVR at various levels of aortic oxygenation.
Results Experimental data were obtained from eight of the 10 newborn lambs. Technical problems early in the development of the animal model associated with the cannulation of the umbilical vessels resulted in prolonged acidosis and death of two of the lambs before arteriovenous support could be initiated. The delivery of arteriovenous flow to the ECMO circuit requires satisfactory myocardial performance. The limiting factors for establishing satisfactory arteriovenous ECMO flows were mean arterial pressure in excess of 30 torr and central venous pressure between 4 and 10 torr. A linear relationship existed between changes in arteriovenous ECMO and total systemic flow rates. The increase in total systemic flow was a reflection of an increase in venous return commensurate with the arteriovenous ECMO rate of perfusion. Similarly, pulmonary and systemic arterial P0 2 varied directly with the arteriovenous rate of perfusion. At the higher arterial P0 2 values (> 50 torr), PBF rose while PVR and the percentage of flow shunted right to left decreased. Arteriovenous ECMO was begun at rates of flow of 20 to 30 ml per kilogram per minute. At these low rates, an initial 7 to 15 torr reduction in peak arterial
pressure from 50 to 65 torr was noted in all experiments. Simultaneously, the central venous pressure fell below 5 torr. Transfusions of 50 to 150 ml. of whole blood raised the central venous pressure to 5 to 10 torr and produced stable peak arterial pressures above 50 torr. Arteriovenous ECMO was continued at these rates of flow for 15 to 30 minutes in order to permit cardiovascular accommodation. Perfusion was subsequently increased incrementally to values approaching 100 to 150 c.c. per kilogram per minute. The variations in aortic and pulmonary arterial P0 2 • total systemic flow (Qt), PBF, and PVR at different rates of arteriovenous flow are shown for a typical experiment in Fig. 3. In the absence of arteriovenous perfusion (Qav = 0), arterial oxygenation was severely depressed and PVR was three times the normal value for lambs at term. As arteriovenous perfusion was increased from zero conditions to rates equalling 20 to 30 percent of the total systemic flow (Qav / Qt = 0.2 to 0.3), pulmonary artery P0 2 rose from base-line values of 17 to 35 torr to 52 to 60 torr. A corresponding rise in P0 2 was measured in the aorta (Table I). The elimination of carbon dioxide was proportional to the arteriovenous ECMO flow: At Qav/ Qt > 0.2, Paco" was reduced by 15 to 20 torr from the base-line conditions in which Qav = 0 (Table I). Right-to-left shunting of blood through the patent foramen ovale and ductus arteriosus varied with arterial oxygenation (Fig. 4). The data compiled from the eight animals of this study were taken for Qav/Qt ratios ranging from zero to 45 percent. A reduction in the percentage of flow shunted right to left under conditions of
Volume 77
Arteriovenous ECMO
Number 4
599
April, 1979
50
I!
p
2,5r-----.-----.-----.-----,
t p<~7' p
10
p
30-38
40-49
50-64
70-112
Pc,02-Torr
Fig. 4. Variation in the right-to-left shuntingof bloodthrough the patent foramen ovale and ductus arteriosus with arterial oxygenation. Results are plotted as the mean ± S.E.M. for data compiled from all animals of this study. hypoxemia (Pao, = 30 to 38 torr) occurred at higher levels of Pao2 (>50 torr) i.e., from 38 percent to 20 to 28 percent. This reduction is significant (p < 0.00 1). An inverse relationship existed between the PVR and level of systemic arterial oxygenation (Fig. 5). PVR was 40 percent lower at Pao2 > 66 torr than at Pao. < 38 torr. The average pulmonary artery pressure was not elevated from prebypass conditions (48/27 torr) at arteriovenous ECMO flows corresponding to less than 30 percent of the total systemic flow (p> 0.05). Discussion Arteriovenous ECMO has been used by others for isolated fetal perfusion. 14-16 Rashkind and associates'? employed arteriovenous perfusion with a bubble oxygenator for gas exchange to support four neonates with hyaline membrane disease and four children with respiratory failure due to cystic fibrosis. Dorson and colleagues" demonstrated the applicability of cannulating the umbilical vessels for arteriovenous ECMO during perfusion of five neonates in respiratory distress. In the present study, experiments were conducted to identify the acute hemodynamic response to arteriovenous perfusion. The results of these acute experiments can be used as a guide for an improved "chronic" mode!. A single distal aortic annul a inserted from an umbilical artery sustained blood flows in excess of 100 m!. per killogram per minute for periods of up to 6 hours. The bypass rates of flow were periodically between 175 and 200 m!. per kilogram per minute. The pressure drop across the polyurethane catheter at flow rates equal to
0-5'----..L..---....I..------'-----' 31-38 40-49 55-64 66-112
PcP2 -Torr
Fig. 5. Reduction in pulmonary vascular resistance with increasinglevelsof arterial oxygenation. The data are plottedas described in Fig. 4. 100 m!. per kilogram per minute was estimated from Poiseuille's law to be 30 torr. Thus the arterial pressure required to establish arteriovenous perfusion at these flows is 30 torr. In the absence of a mechanical pump, this value would be increased to 40 to 55 torr in order to include the low resistance to flow in the microchannel oxygenator (10 to 25 torr). The rate of oxygen transferred by the microchannel oxygenator at arteriovenous ECMO flows between 20 and 45 percent of the total systemic flow ranged between 1.4 and 4.5 C.c. of oxygen per minute (Table I). This figure compares with the oxygen requirement for lambs at 130 days of gestation of 18 c.c. per minute for normothermic conditions. 14 For lambs in the present study which were paralyzed and hypothermic the rate of oxygen delivery in the extracorporeal circuit was sufficient to increase arterial oxygenation to Iifesustaining levels. The therapeutic effects of delivering highly oxygenated blood to the pulmonary vasculature include a reduction in the percentage of flow shunted right to left, an elevation in PBF, and a fall in PVR (Figs. 4 and 5). The fall in PVR with an elevation of PBF suggests that there occurs a vasodilatory effect at the precapillary sphincter mediated by elevation in pulmonary arterial oxygenation." When pulmonary and systemic arterial pressures are the same, the ratio of flow through the ductus to that in the pulmonary circulation varies directly with the systemic vascular resistance and PVR. In neonates with elevated pulmonary vascular resistance and equal pulmonary and systemic arterial pressures, a fetal pattern of circulation exists, characterized by marked right-to-Ieft shunting of blood
The Journal of
6 0 0 Griffith et al.
through the ductus, diminished PBF and consequential systemic hypoxia. The partial respiratory support provided by arteriovenous ECMO causes a therapeutic reversal in this circulatory pattern through delivery of sufficient increments of oxygen to the pulmonary vasculature. Zapol and Kolobow!" presented cineangiographic evidence that the diameter of the ductus arteriosus narrowed markedly at arterial P0 2 values above 35 torr. This implies that the reduction in right-to-left shunting observed in the present study at Pao. above 50 torr may be in part a consequence of the narrowing of the ductus. The reversal in fetal pattern rapidly occurred following the initiation of extracorporeal support and was dependent upon the continuation of ECMO (Fig. 3). The use of arteriovenous ECMO has been restricted due to theories that an increase in venous return leads to an increase in cardiac output, which is deleterious to the myocardium and especially to the right ventricle should pulmonary hypertension exist. 2 However, there is evidence to suggest that partial respiratory support with arteriovenous ECMO can be tolerated by the myocardium of a newborn infant. Rudolph 18 has described the primary change in the circulation of newborn infants at birth as a shift in the flow of blood for gas exchange from the placenta to the lungs. He found right ventricular output in utero to be 300 ml. per kilogram per minute, with 200 rnl. of the flow through the descending aorta being delivered to the placenta. Separation of the infant from the umbilical-placental circulation results in a rapid decrease in venous return and right ventricular output by 200 m!. per kilogram per minute. Postnatally satisfactory oxygenation can be provided by arteriovenous ECMO at rates of flow (Qav < 100 ml, per kilogram per minute) that do not exceed those normally diverted to the placenta. Consequently, right ventricular output should not exceed the flow which previously existed in utero. The results of the present study and studies reported previously": 20 suggest that an initial rise in cardiac output occurs during arteriovenous perfusion in direct response to an increase in the venous return (Fig. 3). This rise appears to be transitory and the cardiac output returns to prebypass levels within I to 2 hours after initiation of arteriovenous ECMO at a constant rate (Qav/Qt :5 0.3). Pulmonary arterial pressure is not elevated and PVR is reduced at those flows which increase arterial oxygenation to life-sustaining levels (Pao. > 50 torr). The lack of variation in electrocardiogram, pulmonary capillary wedge, and central venous pressures indicates that myocardial performance is not impaired.
Thoracic and Cardiovascular Surgery
Partial respiratory support provided by arteriovenous ECMO with the microchannel oxygenator appears to offer special therapeutic advantages in neonates with severe respiratory disease resulting in a fetal circulatory pattern. We wish to acknowledge the design concept for the microchannel oxygenator by Michael H. Weissman, M.D.. Ph.D., in 1969 at Carnegie-Mellon University and to thank Frank McSteen, B.S., for assistance in conducting the animal experiments. REFERENCES Gazzaniga AB, Fong SW, Jefferies MR, Roohk HV, Haiduc M, Bartlett RH: Extracorporeal membrane oxygenator support for cardiopulmonary failure. J THORAC CARDIOVASC SURG 73:375-386, 1976 2 Bartlett RH, Gazzaniga AB: Extracorporeal circulation for cardiopulmonary failure. Curr Probl Surg 15: 1-96. 1978 3 Bohr OF: The pulmonary hypoxic response. Chest 71:244-246, 1977 4 Fishman AP: Hypoxia on the pulmonary circulation. How and where it acts. Circ Res 38: 221-231, 1976 5 Detar R, Gella M: Oxygen and isolated vascular smooth muscle from the main pulmonary artery of the rabbit. Am J Physiol 221:1791-1794, 1971 6 Dorson W, Heyer B, Baker E, Cohen M, Elgis R, Molthan M, Trump 0: Response of distressed infants to partial bypass lung assist. Trans Am Soc Artif Intern Organs 16:345-361, 1970 7 Kolobow T, Zapol W: A new thin-walled non-kinking catheter for peripheral vascular cannulation. Surgery 68:625-629, 1970 8 Hung T-K, Borovetz HS, Hardesty RL, Weissman MH: Toward the development of a neonatal pulmonary assist membrane oxygenator. S Afr Mech Engin 28: 112-116, 1978 9 Hung T-K, Borovetz HS, Weissman MH: Transport and flow phenomena in a microchannel membrane oxygenator. Ann Biomed Engin 5:343-361, 1977 10. Borovetz HS, Griffith BP, Phillips LV, Haubold AD, Hercules, OM, Hung T-K, Hardesty RL: Scanning electron microscopic and surface analytic study of an isotropic vacuum vapor deposited carbon on microporous membranes. Scanning Electron Microosc 11:85-94, 1978 II Griffith BP, Borovetz HS, Hung T-K, Hardesty RL: Continuous indocyanine green dye determination of cardiac output through an arteriovenous shunt. Proc Assoc Advancement Med Inst, Thirteenth Annual Meeting, p 109, 1978 12 Rudolph AM: Cardiac catheterization and angiocardiography, Congenital Diseases of the Heart, Chicago, 1974, Year Book Medical Publishers, Inc., pp 139-149 13 Guyton AC: Textbook of Medical Physiology, Philadelphia, 1971, W. B. Saunders Company, p 206
Volume
rt
Arteriovenous ECMO
Number 4
60 1
April, 1979
14 Zapol W, Kolobow T: Isolated extracorporeal placentation of the fetal lamb, Mammalian Fetus in vitro, CR Austin, ed., London, 1973, Chapman & Hall, Ltd., pp 147-193 15 Doppman JL, Zapol W, Kolobow T, Pierce J: Angiocardiography of fetal lambs on artificial placentas. Invest Radiol 5:181-186, 1970 16 Alexander DP, Britton HG, Nixon DA: Survival of the foetal sheep at term following short periods of perfusion through the umbilical vessels. J Physiol (London) 175:113-124, 1964 17 Rashkind WJ, Milles WW, Falcone D, Taft RW: Hemodynamic effects of arteriovenous oxygenation with a
small volume artificial extracorporeal lung. J Pediatr 70:425-429, 1967 18 Rudolph AM: Changes in the circulation after birth, Congenital Diseases of the Heart, New York, 1974, Year Book Medical Publishers, Inc., pp 17-28 19 Awad JA, Morin PJ: Arteriovenous partial perfusion, Artificial Lungs for Acute Respiratory Failure, WM Zapol, J Qvist, eds., New York, 1976, Academic Press, Inc., pp 329-342 20 Borovetz HS, Griffith BP, Hung T-K, Hardesty RL, Weissman MH: Arteriovenous perfusion with the pulmonary assist membrane oxygenator. Int J Artif Organs 1:232-238, 1978
Information for authors Most of the provisions of the Copyright Act of 1976 became effective on January I, 1978. Therefore, all manuscripts must be accompanied by the following written statement, signed by one author: "The undersigned author transfers all copyright ownership of the manuscript (title of article) to The C. V. Mosby Company in the event the work is published. The undersigned author warrants that the article is original, is not under consideration by another journal, and has not been previously published. I sign for and accept responsibility for releasing this material on behalf of any and all co-authors." Authors will be consulted, when possible, regarding republication of their material.