Development of a Practical Membrane Lung System W. C. Helton, M.D., F. W. Johnson, M.D., J. B. Howe, F. B. Freedman, Ph.D., W. G. Lindsay, M.D., a n d D. M. Nicoloff, M.D., Ph.D. ABSTRACT The purpose of this study was to evaluate a compact, single-pump membrane oxygenator system designed for both short-term open-heart procedures and long-term extracorporeal support. Work was directed at developing a compatible, disposable venous reservoir, fabricating a compact mounting board, determining acceptable priming volumes, and establishing safe and effective modes of setup and operation. Ten adult baboons undergoing total cardiopulmonary autotransplantation as part of a separate study were operated on using a 1.5 mz Kolobow spiral-coil membrane system. Blood gases, oxygen transfer, extracorporeal blood flows, priming volume, and setup times were determined. In addition, 3 infant and 6 adult human subjects undergoing cardiac operations in which spiral-coil membranes of various sizes were used were also similarly studied. The data and experience from these studies suggest that the spiral-coil membrane system utilized is safe and effective in procedures requiring short-term cardiopulmonary bypass.
The availability of a practical membrane oxygenator system for both prolonged cardiopulmonary support and cardiac procedures requiring short-term cardiopulmonary bypass will allow more widespread clinical assessment and utilization of this type of artificial lung. Membrane oxygenators interpose a gas-permeable membrane between the blood and gas phases with gas exchange occurring across the membrane. The advantages of membrane lungs have been reported by others [l,2, 7-91. We have applied this clinical tool for long-term extracorporeal support to patients with cardiac and respiratory failure [5]. Our previous exFrom the Department of Surgery, University of Minnesota Hospitals, and Sci-Med Life Systems, Inc, Minneapolis, MN. We wish to acknowledge the assistance of Mrs. Kathy Mottl and Mr. Don Theis. Accepted for publication Dec 19, 1977. Address reprint requests to Dr. Nicoloff, Box 280, University Hospitals, 412 Union St, SE, Minneapolis, MN 55455.
54 0003-497517810026-0111$01.25@ 1978 by W. C. Helton
perience with this membrane oxygenator demonstrated that patients can be supported with normal blood gases, minimal protein denaturation, thrombocytolysis, and hemolysis for periods of one to twenty days. It is therefore conceivable that such a system should offer some advantage over the current bubble oxygenator, particularly in procedures that can be predicted to require 3 or more hours of bypass. The Kolobow spiral-coil membrane lung [3] was selected for use in conjunction with a series of experimental surgical procedures (total cardiopulmonary autotransplantation) with adult baboons. No attempts were made to evaluate membrane functions critically since this has been done previously [3, 4, 6, 81. Rather, the emphasis was directed at developing a compatible disposable venous reservoir, fabricating a compact mounting board, determining acceptable priming volumes, and establishing safe and effective modes of setup and operation. Emphasis was placed on the areas relating to open-heart surgery. A series of patients was used to clinically evaluate the system based on the success of the laboratory work.
Materials and Methods Bypass Circuit The bypass circuit is illustrated in Figure 1.The membrane lung, heat exchanger, cardiotomy reservoir, and venous reservoir were assembled on a Plexiglas mounting board (Fig 2).* Venous blood was allowed to drain by gravity flow into a silicone rubber reservoir with a capacity of either 500 or 1,500 ml (Fig 3). These reservoirs were constructed so as to accommodate suctioned blood from a cardiotomy reservoir as well as systemic venous return. A vent tube for purging air from the reservoir and two infusion-withdrawal sites were placed at the top of the reservoir. Blood was drawn from the *Travenol heat exchanger, Bentley cardiotomy reservoir, Sci-Med mounting board.
55 Helton et al: Practical Membrane Lung System
Fig I. Bypass circuit: venous blood and cardiotomy suction return to a silicone reservoir. Roller pump draws blood into a heat exchanger and pumps blood through a membrane oxygenator.
venous reservoir through a disposable heat exchanger to the roller pump and then pumped through the membrane to the arterial cannula (Fig 4). A recirculation line was used during priming so that blood flow through the oxygenator could be maintained at all times after the oxygenator had been primed. Recirculation prevented stagnation of formed blood elements within the blood compartment during periods when perfusion was not in progress. The recirculation lines served also as an access site for arterial blood sampling. Cross-matched donor blood (as available) and lactated Ringer's solution were used to prime the circuit. Filtered carbon dioxide was flushed through the blood compartment for 5 minutes prior to priming as an aid in debub-
bling the blood compartment. The gas compartment was closed during this time to maintain the carbon dioxide environment. Blood prime was filtered with a micropore filter when added to the reservoir, and lactated Ringer's solution was added with standard intravenous fluid filtration. Following priming, membrane ventilation (room air suctioned through the gas compartment) and recirculation was done with blood flows maintained at 1 liter per minute. Recirculation with ventilation established approximate physiological pH and Pcoz levels and provided time to debubble the membrane lung. Debubbling was accomplished by tapping the unit gently, but firmly, during recirculation. Positive-pressure membrane ventilation at 4.5 to 7.5 Llmin with either 100°/~oxygen or a 97 to 3 ratio of oxygen to carbon dioxide was used during bypass. The setup time, priming volumes, and ease of assembly were determined for each run.
56 The Annals of Thoracic Surgery Vol 26 No 1 July 1978
Fig 2 . Membrane lun,p system f o r open-heart surgery.
Animal Experience Ten adult baboons ranging from 14 to 18.5 kg were anesthetized with Sernylan (phencyclidine hydrochloride) (2 mg per kilogram of body weight) and atropine sulfate (0.4 mg), intubated, and placed on pressure-limited respirators. Nembutal (pentobarbital) and Flaxedil (gallimine) were given during the procedure to maintain an adequate level of general anesthesia. A right thoracotomy and a femoral artery and venous cutdown were done. The animal was then heparinized with 2 mg per kilogram of beef lung heparin, and the inferior and superior vena cava were cannulated with the largest possible cannula (18 to 24F). Arterial blood was returned through the femoral artery cannula. The baboons were initially placed on total cardiopulmonary bypass and maintained at normothermia for 5 minutes during which time
blood gases were sampled. They were then cooled to 29" to 31°C for the duration of the autotransplant. After the procedure, they were rewarmed to 37°C and allowed to stabilize on partial cardiopulmonary bypass. After stabilization, a second period of normothermic total cardiopulmonary bypass for blood sampling was done. Subsequent to this, cardiopulmonary bypass was discontinued, protamine sulfate (1.5 mg per milligram of heparin) was given to reverse the heparin effect, and the operation was completed. During the course of the experiment, three blood gas samples were drawn: (1) at normothermia after an initial 5-minute period of total bypass (36.5" to 38.0"C); (2) within 5 minutes of reaching the hypothermic state (29" to 31°C); and (3) after a 5-minute period of total bypass at normothermia, just prior to terminating bypass. All samples were drawn distal to the heat exchanger in capped plastic syringes, and the determination was done immediately after the sample was drawn. An Instrumenta-
57 Helton et al: Practical Membrane Lung System
diopulmonary bypass. There were 3 infants weighing 3.2, 3.5, and 3.6 kg and 6 adults weighing 47, 48, 54, 70, and 75 kg, and they ranged from thirty days to 75 years old. tion Laboratory Model 713 blood gas analyzer The bypass circuit, setup, priming techwas used for all blood gas determinations. All niques, and lung ventilation and perfusion temperatures refer to esophageal measure- techniques were identical to those used in the ments. A portion of each sample was used to animal experiments. TWO sizes of venous reserdetermine the hematocrit. Oxygen transfer was voirs were employed depending on whether an calculated from the saturation difference, de- adult or child was being perfused. Membrane rived hemoglobin, and dissolved oxygen. lung size was varied according to the size of the patient and estimated necessary bypass flow. Clinical Experience The flows were estimated at 75 to 100 mUkglmin The clinical experience was evaluated in 9 pa- in the adults and 100 to 150 mUkglmin in infants tients undergoing procedures requiring car- at normothermia, with the membrane able to Fig 3 . Silicone rubber reservoir bags: (A) infant (220 to 500 ml) and (B) adult (800 to 1,500 ml).
58 The Annals of Thoracic Surgery Vol 26 No 1 July 1978
Fi,g 4. Membrane lung system f o r prolonged extracorporeal support. Note the small size of the reservoir and the microswitch control for the roller p u m p .
gases were sampled 5 minutes after instituting total bypass and at the end of 1 hour of total bypass.
carry approximately 1.5 Llm'lmin. Membranes ranging from 0.8 to 4.5 m2 are available and were used.* Cardiotomy in-line filters were used in 4 patients. During priming, hemodilution (16 to 20 mi per kilogram of body weight) was used to adjust the hematocrit to 28 to 30%. Patients were placed on cardiopulmonary bypass and cooled immediately to 29" to 31°C. They were maintained at this level during the cardiac procedure and then rewarmed to 37°C before bypass was discontinued. All blood samples were collected in capped plastic syringes. Determinations were done on an Instrumentation Laboratory Model 713 blood gas analyzer immediately after sampling and corrected for temperature difference. Gas flow rates in the infants ranged from 1.2 to 1 Llmin, and in the adults from 3 to 15 Llmin. Blood
Results Animal Experiments The membrane lung system was evaluated with respect to priming volume, setup time, ease of assembly, operation with closed or open reservoir, and ability to transfer oxygen. The priming volume ranged from 680 to 1,000 ml depending on the reservoir level desired. The mean priming volume was 800 ml with at least 200 ml initially present in the reservoir for all animals. The reservoir functioned effectively with the vents open or closed. However, closed operation was preferred to minimize the risk of drawing air into the bypass circuit. Air accumulation through cardiotomy suction was vented through the venous reservoir bleed ports. Setup time from beginning of assembly to operational status ranged from 20 to 45 minutes (average, 30 min). There were no mechanical problems with any of the membrane lungs. Bypass times
*Membranes of 0.4, 0.8, 1.5, 2.5, 3.5, and 4.5 mz are available from Sci-Med, Inc.
59 Helton et al: Practical Membrane Lung System
Table 1 . Blood Gas Values and Associated Oxygen Transfer Rates Observed for Adult Baboons during Total Bypass (Average and Range) Factor
Blood flow rate (mumidkg body weight) Venous gas Po, (mm Hg) PH Pcoz (mm Hg) Saturation (%) Arterial gas Po, (mg Hg) PH Pcoz (mm Hg) Saturation (%) Oxygen transfer (ccimin) Hematocrit (%) Temperature (“C)
Initial Normothermic State
Hypothermic State, Initial 5 Min
Final Normothermic State
64 (47-82)
46 (35-65)
53 (32-71)
33 (25-37) 7.45 (7.35-7.61) 47 (29-72) 64 (49-78)
23 (15-29) 7.49 (7.43-7.64) 34 (28-40) 69 (39-85)
29 (20-36) 7.36 (7.19-7.47) 50 (34-60) 52 (19-70)
247 (93-375) 7.47 (7.22-7.52) 40 (27-54) 99 (97-100) 50 (24-84) 25 (10-33) 37 (36.5-38.0)
326 (164-296) 7.55 (7.48-7.71) 28 (24-33) 100 (100-100) 34 (14-60) 28 (10-30) 30 (29.0-30.5)
285 (150-410) 7.41 (7.22-7.52) 40 (27-54) 100 (99-100) 52 (24-85)
Table 2 . Blood Gas Values and Oxygen Transfer Rates Observed for Clinical Patients during Total Bypass (Average and Range) Factor Blood flow rate (mllminlkg) Venous gas Po, (mm Hg) Pco, (mm Hg) PH Saturation (%) Arterial gas Po, (mm Hg) Pco, (mm Hg) PH Saturation (YO) Oxygen transfer (ccimin) Hematocrit (YO) Temperature (“C)
Initial Hypothermia 59 (52-63)
1 Hour
59 (48-71)
48 (37-59) 32 (24-41) 7.49 (7.38-7.53) 93 (90-96)
30 (25-35) 31 (26-36) 7.39 (7.34-7.43) 83 (76-96)
263 (181-366) 27 (23-38) 7.45 (7.38-7.50) 98.5 (98-99) 74 (64-82)
150 (108-167) 29 (24-39) 7.41 (7.34-7.47) 98.4 (97-99) 130 (77-205)
34 (28-53) 32 (31-34)
34 (29-42) 31 (28-33)
ranged from 60 to 130 minutes depending on the difficulty of the surgical procedure. All animals survived from 6 hours to 1 year. Four animals died in the first 24 hours due to technical problems during the operation which resulted in persistent bleeding from anastomotic sites. Five animals died at one to six days from various causes, including progressive arrhythmias, cardiorespiratory failure, and right
23 (10-34) 36.7 (36.0-37.5)
Table 3 . Average Platelet Counts for Clinical Patients Undergoing Bypass Time
Count
Preoperative Initial 5 min of bypass After 1 hr of bypass One hr postoperative
182,000 (118-225 144,000 (111-179 155,000 (108-202 178,000 (156-196
X X
X X
lo3) 103) lo3) lo3)
atrial thrombi at the suture line resulting in pulmonary embolization. At the time of writing, there was 1 long-term survivor from this series of animals. Blood flows, blood gases, and oxygen transfer rates are shown in Table 1.
Clinical Experience Priming volume of the infant unit was approximately 1,000 ml. The adult unit required an average of 2,000 ml priming volume (range, 1,700 to 2,300 ml). The setup time from unpacking the components to the operational status averaged 25 minutes (range, 15 to 45 min). There were no mechanical problems during the bypass period. The average blood gas values and oxygen transfer rates for adults are presented in Table 2. A summary of platelet counts done before, during, and after bypass for adults is shown in Table 3. A representative data sheet
60 The Annals of Thoracic Surgery Vol 26 No 1 July 1978
Table 4. Data on 70-kg Patient Obtained Utilizing a 3.5 m2 Kolobow Membrane Lung Bypass Time Data
5 min
60 min
Arterial blood pressure (mm Hg) Extracorporeal blood flow (Llmin) Arterial values Po2 (mm Hg) PH Pcoz (mm Hg) Saturation (O/O) Venous values Po, (mm Hg) PH Pco2(mm Hg) Saturation (O/O) Rectal temperature ("C) Esophageal temperature ("C) Oxygen flow rate (Llmin) Oxygen transfer (cclmin) Potassium (mEqlL) Sodium (mEqlL) Platelets (per mm3) Hematocrit (YO) Total urine (ml)
58
100
4.4
5.0
250 7.44 25 98.5
167 7.47 24 99
37 7.53 30.5 90 33 31 4.5 75 3.3 124 111,000 28 ...
25 7.43 30 78 31 31 3.0 134 4.4 125 108,000 28.5 400
from a single patient is demonstrated in Table 4.
Comment The membrane lung system was evaluated experimentally and clinically with regard to the logistics of assembly and operation. Membrane function per se has been previously evaluated [3,4, 61 and therefore was not rigorously determined in this study. Emphasis was directed at the perfusion circuit design, fabrication of component parts, and the requirements for a practical system. The position of the circuit components and setup techniques were developed in a series of preliminary studies with dogs placed on cardiopulmonary bypass. The venous reservoir was placed before the heat exchanger to facilitate bubble removal from the venous lines. The geometry of the reservoir (see Fig 2) was such that air was easily removed, with stagnation and turbulence being minimal. The membrane
was placed at 30 degrees from the vertical with the upper bleed port (Luer-Lok) at the highest point to aid in bubble removal from the membrane itself. The Plexiglas mounting board was developed to facilitate assembly of the circuit components. The mounting lugs for the circuit component are adjustable in order to accommodate the variously sized membranes, reservoirs, and heat exchangers. Prior to priming, positive-pressure ventilation of the blood compartment with carbon dioxide was done to aid in bubble removal. A gentle tapping of the membrane with a soft mallet during priming caused the bubbles to rise to the surface, where they were removed through the upper bleed port. Positive gas pressures were used during bypass rather than vacuum ventilation. This was done since positive pressure increases oxygen transfer compared with vacuum ventilation. There are two reasons for this: positive pressure creates a higher driving force for transfer because of the higher Po, in the gas compartment, and it decreases the blood channel thickness because of lower transmembrane pressure. Gas embolization was prevented since the positive pressure within the blood compartment of the membrane was always greater than the pressure within the gas compartment. A pressure plate was placed over the venous reservoir bag and calibrated in relation to reservoir volume. The animal data with regard to the logistics of assembly and operation were considered adequate to warrant a clinical trial. However, gas exchange data were marginally acceptable even though the membranes were of adequate size for hypothermia and a brief period of normothermia. Work by Rawitscher, Dutton, and Edmunds [6] indicates that the oxygen transfer capacity of the 1.5 m2 membrane is approximately 75 cclmin at rated flow conditions (hematocrit, 40%; temperature, 37°C; inflow saturation, 65%; and blood flow at 1.0 L/m2/ min). In our experience under these conditions but with a blood flow of 1.5 Llm', oxygen transfer of approximately 100 cclmin can be obtained. Since the oxygen requirements for these animals range from 100 to 125 cclmin at normothermia, it was considered that the 1.5 m2 membrane would be adequate for a hypothermic bypass with only a brief period of nor-
61 Helton et al: Practical Membrane Lung System
mothermia. The low venous Po, a n d metabolic acidosis during bypass are reflections of a progressive oxygen d e b t occurring during bypass. This was thought to be a function of low bypass hematocrits a n d low blood flow rates. These low hematocrits a n d blood flow rates were tolerated since the baboon colony w a s of limited size a n d t h e only source of donor blood. The donor blood h a d to be carefully proportioned between priming requirements a n d postoperative transfusion needs. Undoubtedly, the progressive acidosis that resulted from the low bypass flows a n d hematocrits w a s a factor in the postoperative death of several of the animals. The blood gas data in the h u m a n series w a s always in the acceptable ranges, as were the blood flow rates a n d hematocrits d u r i n g bypass. The initial decrease in platelet counts during bypass w a s due to hernodilution. Subsequent platelet counts were slightly lower t h a n could be explained by dilution alone. No significant postoperative bleeding w a s demonstrated. The oxygen transfer rates in t h e adults ranged from 73 to 235 cclmin. The lower values for oxygen transfer in this group of patients w a s thought to be a function of t h e high venous saturation. This, in turn, w a s considered a reflection of the low oxygen d e m a n d s of hypothermia and some element of peripheral shunting.
References 1. Chopra PS, Dufek JH, Kroncke GM, et al: Clinical
2.
3. 4.
5. 6.
7.
8.
9.
comparison of the General Electric-Peirce membrane lung and bubble oxygenator for prolonged cardiopulmonary bypass. Surgery 745374, 1973 Dutton RC, Edmunds LH Jr, Hutchinson JC, et al: Platelet aggregate emboli produced in patients during cardiopulmonary bypass with membrane and bubble oxygenators and blood filters. J Thorac Cardiovasc Surg 67:258, 1974 Kolobow T, Bowman RL: Construction and evaluation of an alveolar membrane artificial heart-lung. Trans Am SOCArtif Intern Organs 9:238, 1963 Murphy WRC, Galleti PM, Richardson PD: Determinants of performance in spiral coil membrane oxygenators. Published in the Third New England Biomedical Engineering Conference. May, 1975 Pyle RB, Helton WC, Johnson RW, et al: Clinical use of the membrane oxygenator. Arch Surg 110:8:966, 1975 Rawitscher RE, Dutton RC, Edmunds LH Jr: Evaluation of hollow fiber and spiral coil membrane oxygenators designed for cardiopulmonary bypass in infants. Circulation 47, 48:Suppl 3105, 1973 Siderys H, Herod GT, Halbrook H, et al: A comparison of membrane and bubble oxygenation as used in cardiopulmonary bypass in patients: the importance of pericardial blood as a source of hemolysis. J Thorac Cardiovasc Surg 69:708, 1975 Ward BD, Hood AG: Lung comparative performance of clinical membrane lungs. J Thorac Cardiovasc Surg 68:830, 1974 Wright JS, Fisk GC, Torda TA, et al: Some advantages of the membrane oxygenator for open heart surgery. J Thorac Cardiovasc Surg 69:884, 1975
The availability of a practical membrane oxygenator system for both prolonged cardiopulmonary support and cardiac procedures requiring short-term cardiopulmonary bypass will allow more widespread clinical assessment and utilization of this type of artificial lung. Membrane oxygenators interpose a gas-permeable membrane between the blood and gas phases with gas exchange occurring across the membrane. The advantages of membrane lungs have been reported by others [l,2, 7-91. We have applied this clinical tool for long-term extracorporeal support to patients with cardiac and respiratory failure [5]. Our previous exFrom the Department of Surgery, University of Minnesota Hospitals, and Sci-Med Life Systems, Inc, Minneapolis, MN. We wish to acknowledge the assistance of Mrs. Kathy Mottl and Mr. Don Theis. Accepted for publication Dec 19, 1977. Address reprint requests to Dr. Nicoloff, Box 280, University Hospitals, 412 Union St, SE, Minneapolis, MN 55455.
54 0003-497517810026-0111$01.25@ 1978 by W. C. Helton
perience with this membrane oxygenator demonstrated that patients can be supported with normal blood gases, minimal protein denaturation, thrombocytolysis, and hemolysis for periods of one to twenty days. It is therefore conceivable that such a system should offer some advantage over the current bubble oxygenator, particularly in procedures that can be predicted to require 3 or more hours of bypass. The Kolobow spiral-coil membrane lung [3] was selected for use in conjunction with a series of experimental surgical procedures (total cardiopulmonary autotransplantation) with adult baboons. No attempts were made to evaluate membrane functions critically since this has been done previously [3, 4, 6, 81. Rather, the emphasis was directed at developing a compatible disposable venous reservoir, fabricating a compact mounting board, determining acceptable priming volumes, and establishing safe and effective modes of setup and operation. Emphasis was placed on the areas relating to open-heart surgery. A series of patients was used to clinically evaluate the system based on the success of the laboratory work.
Materials and Methods Bypass Circuit The bypass circuit is illustrated in Figure 1.The membrane lung, heat exchanger, cardiotomy reservoir, and venous reservoir were assembled on a Plexiglas mounting board (Fig 2).* Venous blood was allowed to drain by gravity flow into a silicone rubber reservoir with a capacity of either 500 or 1,500 ml (Fig 3). These reservoirs were constructed so as to accommodate suctioned blood from a cardiotomy reservoir as well as systemic venous return. A vent tube for purging air from the reservoir and two infusion-withdrawal sites were placed at the top of the reservoir. Blood was drawn from the *Travenol heat exchanger, Bentley cardiotomy reservoir, Sci-Med mounting board.
55 Helton et al: Practical Membrane Lung System
Fig I. Bypass circuit: venous blood and cardiotomy suction return to a silicone reservoir. Roller pump draws blood into a heat exchanger and pumps blood through a membrane oxygenator.
venous reservoir through a disposable heat exchanger to the roller pump and then pumped through the membrane to the arterial cannula (Fig 4). A recirculation line was used during priming so that blood flow through the oxygenator could be maintained at all times after the oxygenator had been primed. Recirculation prevented stagnation of formed blood elements within the blood compartment during periods when perfusion was not in progress. The recirculation lines served also as an access site for arterial blood sampling. Cross-matched donor blood (as available) and lactated Ringer's solution were used to prime the circuit. Filtered carbon dioxide was flushed through the blood compartment for 5 minutes prior to priming as an aid in debub-
bling the blood compartment. The gas compartment was closed during this time to maintain the carbon dioxide environment. Blood prime was filtered with a micropore filter when added to the reservoir, and lactated Ringer's solution was added with standard intravenous fluid filtration. Following priming, membrane ventilation (room air suctioned through the gas compartment) and recirculation was done with blood flows maintained at 1 liter per minute. Recirculation with ventilation established approximate physiological pH and Pcoz levels and provided time to debubble the membrane lung. Debubbling was accomplished by tapping the unit gently, but firmly, during recirculation. Positive-pressure membrane ventilation at 4.5 to 7.5 Llmin with either 100°/~oxygen or a 97 to 3 ratio of oxygen to carbon dioxide was used during bypass. The setup time, priming volumes, and ease of assembly were determined for each run.
56 The Annals of Thoracic Surgery Vol 26 No 1 July 1978
Fig 2 . Membrane lun,p system f o r open-heart surgery.
Animal Experience Ten adult baboons ranging from 14 to 18.5 kg were anesthetized with Sernylan (phencyclidine hydrochloride) (2 mg per kilogram of body weight) and atropine sulfate (0.4 mg), intubated, and placed on pressure-limited respirators. Nembutal (pentobarbital) and Flaxedil (gallimine) were given during the procedure to maintain an adequate level of general anesthesia. A right thoracotomy and a femoral artery and venous cutdown were done. The animal was then heparinized with 2 mg per kilogram of beef lung heparin, and the inferior and superior vena cava were cannulated with the largest possible cannula (18 to 24F). Arterial blood was returned through the femoral artery cannula. The baboons were initially placed on total cardiopulmonary bypass and maintained at normothermia for 5 minutes during which time
blood gases were sampled. They were then cooled to 29" to 31°C for the duration of the autotransplant. After the procedure, they were rewarmed to 37°C and allowed to stabilize on partial cardiopulmonary bypass. After stabilization, a second period of normothermic total cardiopulmonary bypass for blood sampling was done. Subsequent to this, cardiopulmonary bypass was discontinued, protamine sulfate (1.5 mg per milligram of heparin) was given to reverse the heparin effect, and the operation was completed. During the course of the experiment, three blood gas samples were drawn: (1) at normothermia after an initial 5-minute period of total bypass (36.5" to 38.0"C); (2) within 5 minutes of reaching the hypothermic state (29" to 31°C); and (3) after a 5-minute period of total bypass at normothermia, just prior to terminating bypass. All samples were drawn distal to the heat exchanger in capped plastic syringes, and the determination was done immediately after the sample was drawn. An Instrumenta-
57 Helton et al: Practical Membrane Lung System
diopulmonary bypass. There were 3 infants weighing 3.2, 3.5, and 3.6 kg and 6 adults weighing 47, 48, 54, 70, and 75 kg, and they ranged from thirty days to 75 years old. tion Laboratory Model 713 blood gas analyzer The bypass circuit, setup, priming techwas used for all blood gas determinations. All niques, and lung ventilation and perfusion temperatures refer to esophageal measure- techniques were identical to those used in the ments. A portion of each sample was used to animal experiments. TWO sizes of venous reserdetermine the hematocrit. Oxygen transfer was voirs were employed depending on whether an calculated from the saturation difference, de- adult or child was being perfused. Membrane rived hemoglobin, and dissolved oxygen. lung size was varied according to the size of the patient and estimated necessary bypass flow. Clinical Experience The flows were estimated at 75 to 100 mUkglmin The clinical experience was evaluated in 9 pa- in the adults and 100 to 150 mUkglmin in infants tients undergoing procedures requiring car- at normothermia, with the membrane able to Fig 3 . Silicone rubber reservoir bags: (A) infant (220 to 500 ml) and (B) adult (800 to 1,500 ml).
58 The Annals of Thoracic Surgery Vol 26 No 1 July 1978
Fi,g 4. Membrane lung system f o r prolonged extracorporeal support. Note the small size of the reservoir and the microswitch control for the roller p u m p .
gases were sampled 5 minutes after instituting total bypass and at the end of 1 hour of total bypass.
carry approximately 1.5 Llm'lmin. Membranes ranging from 0.8 to 4.5 m2 are available and were used.* Cardiotomy in-line filters were used in 4 patients. During priming, hemodilution (16 to 20 mi per kilogram of body weight) was used to adjust the hematocrit to 28 to 30%. Patients were placed on cardiopulmonary bypass and cooled immediately to 29" to 31°C. They were maintained at this level during the cardiac procedure and then rewarmed to 37°C before bypass was discontinued. All blood samples were collected in capped plastic syringes. Determinations were done on an Instrumentation Laboratory Model 713 blood gas analyzer immediately after sampling and corrected for temperature difference. Gas flow rates in the infants ranged from 1.2 to 1 Llmin, and in the adults from 3 to 15 Llmin. Blood
Results Animal Experiments The membrane lung system was evaluated with respect to priming volume, setup time, ease of assembly, operation with closed or open reservoir, and ability to transfer oxygen. The priming volume ranged from 680 to 1,000 ml depending on the reservoir level desired. The mean priming volume was 800 ml with at least 200 ml initially present in the reservoir for all animals. The reservoir functioned effectively with the vents open or closed. However, closed operation was preferred to minimize the risk of drawing air into the bypass circuit. Air accumulation through cardiotomy suction was vented through the venous reservoir bleed ports. Setup time from beginning of assembly to operational status ranged from 20 to 45 minutes (average, 30 min). There were no mechanical problems with any of the membrane lungs. Bypass times
*Membranes of 0.4, 0.8, 1.5, 2.5, 3.5, and 4.5 mz are available from Sci-Med, Inc.
59 Helton et al: Practical Membrane Lung System
Table 1 . Blood Gas Values and Associated Oxygen Transfer Rates Observed for Adult Baboons during Total Bypass (Average and Range) Factor
Blood flow rate (mumidkg body weight) Venous gas Po, (mm Hg) PH Pcoz (mm Hg) Saturation (%) Arterial gas Po, (mg Hg) PH Pcoz (mm Hg) Saturation (%) Oxygen transfer (ccimin) Hematocrit (%) Temperature (“C)
Initial Normothermic State
Hypothermic State, Initial 5 Min
Final Normothermic State
64 (47-82)
46 (35-65)
53 (32-71)
33 (25-37) 7.45 (7.35-7.61) 47 (29-72) 64 (49-78)
23 (15-29) 7.49 (7.43-7.64) 34 (28-40) 69 (39-85)
29 (20-36) 7.36 (7.19-7.47) 50 (34-60) 52 (19-70)
247 (93-375) 7.47 (7.22-7.52) 40 (27-54) 99 (97-100) 50 (24-84) 25 (10-33) 37 (36.5-38.0)
326 (164-296) 7.55 (7.48-7.71) 28 (24-33) 100 (100-100) 34 (14-60) 28 (10-30) 30 (29.0-30.5)
285 (150-410) 7.41 (7.22-7.52) 40 (27-54) 100 (99-100) 52 (24-85)
Table 2 . Blood Gas Values and Oxygen Transfer Rates Observed for Clinical Patients during Total Bypass (Average and Range) Factor Blood flow rate (mllminlkg) Venous gas Po, (mm Hg) Pco, (mm Hg) PH Saturation (%) Arterial gas Po, (mm Hg) Pco, (mm Hg) PH Saturation (YO) Oxygen transfer (ccimin) Hematocrit (YO) Temperature (“C)
Initial Hypothermia 59 (52-63)
1 Hour
59 (48-71)
48 (37-59) 32 (24-41) 7.49 (7.38-7.53) 93 (90-96)
30 (25-35) 31 (26-36) 7.39 (7.34-7.43) 83 (76-96)
263 (181-366) 27 (23-38) 7.45 (7.38-7.50) 98.5 (98-99) 74 (64-82)
150 (108-167) 29 (24-39) 7.41 (7.34-7.47) 98.4 (97-99) 130 (77-205)
34 (28-53) 32 (31-34)
34 (29-42) 31 (28-33)
ranged from 60 to 130 minutes depending on the difficulty of the surgical procedure. All animals survived from 6 hours to 1 year. Four animals died in the first 24 hours due to technical problems during the operation which resulted in persistent bleeding from anastomotic sites. Five animals died at one to six days from various causes, including progressive arrhythmias, cardiorespiratory failure, and right
23 (10-34) 36.7 (36.0-37.5)
Table 3 . Average Platelet Counts for Clinical Patients Undergoing Bypass Time
Count
Preoperative Initial 5 min of bypass After 1 hr of bypass One hr postoperative
182,000 (118-225 144,000 (111-179 155,000 (108-202 178,000 (156-196
X X
X X
lo3) 103) lo3) lo3)
atrial thrombi at the suture line resulting in pulmonary embolization. At the time of writing, there was 1 long-term survivor from this series of animals. Blood flows, blood gases, and oxygen transfer rates are shown in Table 1.
Clinical Experience Priming volume of the infant unit was approximately 1,000 ml. The adult unit required an average of 2,000 ml priming volume (range, 1,700 to 2,300 ml). The setup time from unpacking the components to the operational status averaged 25 minutes (range, 15 to 45 min). There were no mechanical problems during the bypass period. The average blood gas values and oxygen transfer rates for adults are presented in Table 2. A summary of platelet counts done before, during, and after bypass for adults is shown in Table 3. A representative data sheet
60 The Annals of Thoracic Surgery Vol 26 No 1 July 1978
Table 4. Data on 70-kg Patient Obtained Utilizing a 3.5 m2 Kolobow Membrane Lung Bypass Time Data
5 min
60 min
Arterial blood pressure (mm Hg) Extracorporeal blood flow (Llmin) Arterial values Po2 (mm Hg) PH Pcoz (mm Hg) Saturation (O/O) Venous values Po, (mm Hg) PH Pco2(mm Hg) Saturation (O/O) Rectal temperature ("C) Esophageal temperature ("C) Oxygen flow rate (Llmin) Oxygen transfer (cclmin) Potassium (mEqlL) Sodium (mEqlL) Platelets (per mm3) Hematocrit (YO) Total urine (ml)
58
100
4.4
5.0
250 7.44 25 98.5
167 7.47 24 99
37 7.53 30.5 90 33 31 4.5 75 3.3 124 111,000 28 ...
25 7.43 30 78 31 31 3.0 134 4.4 125 108,000 28.5 400
from a single patient is demonstrated in Table 4.
Comment The membrane lung system was evaluated experimentally and clinically with regard to the logistics of assembly and operation. Membrane function per se has been previously evaluated [3,4, 61 and therefore was not rigorously determined in this study. Emphasis was directed at the perfusion circuit design, fabrication of component parts, and the requirements for a practical system. The position of the circuit components and setup techniques were developed in a series of preliminary studies with dogs placed on cardiopulmonary bypass. The venous reservoir was placed before the heat exchanger to facilitate bubble removal from the venous lines. The geometry of the reservoir (see Fig 2) was such that air was easily removed, with stagnation and turbulence being minimal. The membrane
was placed at 30 degrees from the vertical with the upper bleed port (Luer-Lok) at the highest point to aid in bubble removal from the membrane itself. The Plexiglas mounting board was developed to facilitate assembly of the circuit components. The mounting lugs for the circuit component are adjustable in order to accommodate the variously sized membranes, reservoirs, and heat exchangers. Prior to priming, positive-pressure ventilation of the blood compartment with carbon dioxide was done to aid in bubble removal. A gentle tapping of the membrane with a soft mallet during priming caused the bubbles to rise to the surface, where they were removed through the upper bleed port. Positive gas pressures were used during bypass rather than vacuum ventilation. This was done since positive pressure increases oxygen transfer compared with vacuum ventilation. There are two reasons for this: positive pressure creates a higher driving force for transfer because of the higher Po, in the gas compartment, and it decreases the blood channel thickness because of lower transmembrane pressure. Gas embolization was prevented since the positive pressure within the blood compartment of the membrane was always greater than the pressure within the gas compartment. A pressure plate was placed over the venous reservoir bag and calibrated in relation to reservoir volume. The animal data with regard to the logistics of assembly and operation were considered adequate to warrant a clinical trial. However, gas exchange data were marginally acceptable even though the membranes were of adequate size for hypothermia and a brief period of normothermia. Work by Rawitscher, Dutton, and Edmunds [6] indicates that the oxygen transfer capacity of the 1.5 m2 membrane is approximately 75 cclmin at rated flow conditions (hematocrit, 40%; temperature, 37°C; inflow saturation, 65%; and blood flow at 1.0 L/m2/ min). In our experience under these conditions but with a blood flow of 1.5 Llm', oxygen transfer of approximately 100 cclmin can be obtained. Since the oxygen requirements for these animals range from 100 to 125 cclmin at normothermia, it was considered that the 1.5 m2 membrane would be adequate for a hypothermic bypass with only a brief period of nor-
61 Helton et al: Practical Membrane Lung System
mothermia. The low venous Po, a n d metabolic acidosis during bypass are reflections of a progressive oxygen d e b t occurring during bypass. This was thought to be a function of low bypass hematocrits a n d low blood flow rates. These low hematocrits a n d blood flow rates were tolerated since the baboon colony w a s of limited size a n d t h e only source of donor blood. The donor blood h a d to be carefully proportioned between priming requirements a n d postoperative transfusion needs. Undoubtedly, the progressive acidosis that resulted from the low bypass flows a n d hematocrits w a s a factor in the postoperative death of several of the animals. The blood gas data in the h u m a n series w a s always in the acceptable ranges, as were the blood flow rates a n d hematocrits d u r i n g bypass. The initial decrease in platelet counts during bypass w a s due to hernodilution. Subsequent platelet counts were slightly lower t h a n could be explained by dilution alone. No significant postoperative bleeding w a s demonstrated. The oxygen transfer rates in t h e adults ranged from 73 to 235 cclmin. The lower values for oxygen transfer in this group of patients w a s thought to be a function of t h e high venous saturation. This, in turn, w a s considered a reflection of the low oxygen d e m a n d s of hypothermia and some element of peripheral shunting.
References 1. Chopra PS, Dufek JH, Kroncke GM, et al: Clinical
2.
3. 4.
5. 6.
7.
8.
9.
comparison of the General Electric-Peirce membrane lung and bubble oxygenator for prolonged cardiopulmonary bypass. Surgery 745374, 1973 Dutton RC, Edmunds LH Jr, Hutchinson JC, et al: Platelet aggregate emboli produced in patients during cardiopulmonary bypass with membrane and bubble oxygenators and blood filters. J Thorac Cardiovasc Surg 67:258, 1974 Kolobow T, Bowman RL: Construction and evaluation of an alveolar membrane artificial heart-lung. Trans Am SOCArtif Intern Organs 9:238, 1963 Murphy WRC, Galleti PM, Richardson PD: Determinants of performance in spiral coil membrane oxygenators. Published in the Third New England Biomedical Engineering Conference. May, 1975 Pyle RB, Helton WC, Johnson RW, et al: Clinical use of the membrane oxygenator. Arch Surg 110:8:966, 1975 Rawitscher RE, Dutton RC, Edmunds LH Jr: Evaluation of hollow fiber and spiral coil membrane oxygenators designed for cardiopulmonary bypass in infants. Circulation 47, 48:Suppl 3105, 1973 Siderys H, Herod GT, Halbrook H, et al: A comparison of membrane and bubble oxygenation as used in cardiopulmonary bypass in patients: the importance of pericardial blood as a source of hemolysis. J Thorac Cardiovasc Surg 69:708, 1975 Ward BD, Hood AG: Lung comparative performance of clinical membrane lungs. J Thorac Cardiovasc Surg 68:830, 1974 Wright JS, Fisk GC, Torda TA, et al: Some advantages of the membrane oxygenator for open heart surgery. J Thorac Cardiovasc Surg 69:884, 1975
