Development of a new hollow fiber silicone membrane oxygenator for ECMO: the recent progress.

Throughout the last 50 years, many improvements have been made for a more effective oxygenator. A large plate type membrane oxygenator, used by Clowes, and a coil type, used by Kolff, has evolved into the small hollow fiber oxygenator. The complex bubble oxygenator, or rotating disk oxygenator, has become a small disposable bubble oxygenator. The currently available oxygenators are extremely small, efficient, and can be used for extended periods of time. However, there are some problems with extracorporeal membrane oxygenation (ECMO). Currently in the United States, there are no clinically applicable hollow fiber ECMO oxygenators available, in spite of the extended ECMO application. Therefore, the development of a small, yet efficient, silicone hollow fiber membrane oxygenator for long-term ECMO usage was attempted. Based on the results of many experimental models, preclinical oxygenator models for long-term ECMO were developed in our laboratory using an ultra-thin silicone rubber hollow fiber membrane.     

Haematological characteristics of a new membrane oxygenator: the Cobe CML

The new Cobe CML membrane oxygenator is more compact than other membrane oxygenators and has a combined venous and cardiotomy suction reservoir. Its size makes it as easy to use as a bubble oxygenator. The studies reported here were designed to show whether the excellent haemocompatibility found with other types of membrane oxygenators had ben compromised by the changes introduced in the Cobe CML oxygenator. Platelet number and function (ADP induced aggregation) plasma betathromboglobulin concentration and plasma haemoglobulin concentration were studied in nine patients where the Cobe CML oxygenator had been used and these were compared with ten patients managed with a Shiley S-100 bubble oxygenator. We conclude that the constructional changes of the Cobe CML oxygenator do not affect the haemocompatibility of this type of membrane oxygenator and that it remains significantly better than the Shiley S-100 bubble oxygenator.     

Membrane oxygenator prevents lung reperfusion injury in canine cardiopulmonary bypass

The effect of blood activation on lung reperfusion injury during cardiopulmonary bypass was investigated in 20 dogs with the use of a bubble oxygenator (n = 10) or a membrane oxygenator (n = 10). In the bubble oxygenator group, significant leukocyte and platelet right to left atrium gradients were found 15 minutes after lung reperfusion (p less than 0.05, p less than 0.01) accompanied by a sharp increase in plasma malondialdehyde concentration 5 minutes after lung reperfusion, whereas no significant right to left atrium gradient of leukocytes or platelets nor significant increase in plasma malondialdehyde concentration was observed in the membrane oxygenator group. In both the bubble oxygenator and membrane oxygenator group, similar mild to moderate lung histological changes were found before lung reperfusion. After lung reperfusion, however, more endothelial cell swelling (p less than 0.05), leukocyte (p less than 0.01) and platelet (p less than 0.01) accumulation in lung capillaries, leakage of erythrocytes into the alveolar space (p less than 0.05), and type I cell damage (p less than 0.05) were found only in the bubble oxygenator group. Eventually, a significantly higher lung water content was found in the bubble oxygenator group than in the membrane oxygenator group (p less than 0.01) after cardiopulmonary bypass. This study indicated that lung injury during cardiopulmonary bypass starts mainly after lung reperfusion, which was correlated with lung leukocyte and platelet sequestration associated with different types of oxygenators.     

From the spinning disc to the membrane oxygenator for open-heart surgery

Gibbon's rotating cylinder could not be enlarged to oxygenate an animal larger than a cat. The spinning disc oxygenator, introduced in 1947, had the capacity to perfuse a dog and the potential to increase oxygenation capacity by addition of more discs. When centers began to do three to four open-heart operations per day, the disposable bubble oxygenator was more practical. Bubble size was optimized to decrease the flow of oxygen relative to the blood flow and reduce trauma to blood. The bubble oxygenator is the type most commonly used today. Use of deep hypothermia with whole blood at an esophageal temperature of 10 degrees C was initially complicated by brain damage due to aggregation of white blood corpuscles and platelets. The introduction of hemodilution permitted safe utilization of hypothermic perfusion. Perfusion of infants should not be carried out at hematocrit below 25 ml/100 m. Early membrane oxygenators used nonporous silicone, or modified silicone membranes. High priming volumes, high pressure drop and marginal gas transfer efficiency characterized these devices. Recent advances in membrane technology have spawned a new generation of membrane oxygenators utilizing microporous polypropylene. In these new oxygenators, with either microporous hollow fibers or sheet membrane, the gas transfer characteristics are far superior to those of types produced in the past. The hollow-fiber devices typically have larger surface areas and higher pressure drop than in the new state-of-the-art flat plate models. An evaluation of one of these new-generation membrane oxygenators gave optimal oxygen and carbon dioxide exchange at a gas flow of 1 l/min of 60% oxygen in air at 30 degrees C and 2 l/min of 80% oxygen in air at normal temperature and rewarming for an adult. Today, after almost 40 years of oxygenator development, these new membrane device can offer better platelet preservation and reduced blood trauma as compared with types developed in the past. The new membrane oxygenators are fast becoming the preferred choice for use in infants and in protracted perfusion.     

New membrane oxygenator (LPM 50): influence on extravascular lung water and pulmonary function in comparison to bubble oxygenator

A prospective, randomized, clinical study involving 30 patients undergoing aorta-coronary bypass grafting was designed to compare the influence of a new membrane oxygenator and a commonly used bubble oxygenator on extravascular lung water and pulmonary function after extracorporeal circulation. Although membrane oxygenators might have some advantages from the biochemical and biophysical points of view, in this clinical study no differences in lung water accumulation and pulmonary gas exchange could be detected between bubble and membrane oxygenators after extracorporeal circulation.     

Retinal microembolism and neuropsychological deficit following clinical cardiopulmonary bypass: comparison of a membrane and a bubble oxygenator. A preliminary communication

To observe and quantify cerebrovascular microembolic events in the central nervous system during cardiopulmonary bypass, 40 patients having elective uncomplicated coronary surgery had retinal fluorescein angiograms 5 min before bypass was discontinued. Each patient also had 10 neuropsychological tests before and after surgery. A Harvey H1700 bubble oxygenator was used for 23 patients and a Cobe CML sheet membrane oxygenator was used for 17 patients. All 23 (100%) of patients in the bubble oxygenator group had retinal microvascular occlusions consistent with microembolism compared to 8/17 (47%) in the membrane oxygenator group (P less than 0.001). In those retinas with occlusions, the mean resultant area of non-perfusion was less in the membrane oxygenator group (0.11 mm2; n = 8) than in the bubble oxygenator group (0.29 mm2; P less than 0.01). Arterial PO2 levels during bypass were similar in both groups at moderate hypothermia, but the mean PaO2 during rewarming was higher in the bubble oxygenator group (27 kPa) than in the membrane group (13 kPa; P less than 0.001). Neuropsychological deficits were more common and more severe after bubble oxygenation than after membrane oxygenation, but in this small patient group, the difference was not statistically significant. We conclude that flat sheet membrane oxygenation during cardiopulmonary bypass may confer significant protection against cerebrovascular microembolism.     

A comparative clinical assessment of a hollow-fibre membrane oxygenator (Capiox II) and a bubble oxygenator (Harvey 1500)

Personal experience is reported on the use of a membrane oxygenator, the Capiox II, which is clinically compared with a bubble oxygenator for medium-term perfusion. The characteristics considered were the efficiency of the heat exchanger, the oxygenating capacity, traumatic effects on the blood and the direct effect on the renal and cardiopulmonary systems. The Capiox II demonstrated a better oxygenating capacity, less platelet damage, a smaller variation in the free plasma haemoglobin, a significant difference in postoperative bleeding and blood transfusion requirements, but showed no variation in renal and cardiopulmonary function. The authors suggest that the Capiox II is to be preferred for extra-corporeal circulation of medium duration, for which it combines the advantages of both bubble and membrane oxygenators, while being less complex and costly than previously marketed membrane devices.     

Clinical experience with the Teflo disposable membrane oxygenator.

A simple, inexpensive, highly efficient disposable membrane oxygenator with low priming volume and a microporous membrane recently has become available. Animal and clinical investigations of its use have been most satisfactory, and clinical experience now has been extended to include 285 patients. Its primary advantage has been the ability to control oxygenation and carbon dioxide separately. Disadvantages have included the somewhat increased complexity of the system as compared with bubble oxygenator systems, the necessity of converting pumps for its use, and excessive condensation of water vapor in the gas phase of the oxygenator unless certain precautions are followed.     

Influence of oxygenator type on the prevalence and extent of microembolic retinal ischemia during cardiopulmonary bypass. Assessment by digital image analysis

We have previously reported the occurrence of microembolic ischemia in the retina during cardiopulmonary bypass, as revealed by fluorescein angiography. This method has been extended by digital image analysis to include quantification of the extent of retinal ischemia and has been applied to a prospective comparative study of 64 patients undergoing elective coronary operations with either a bubble or a membrane oxygenator. Patients with diabetes or clinically evident cerebrovascular disease were excluded. Bypass procedures were standardized in all cases with pulsatile flow and a 40 microns arterial line filter (Pall EC Plus). Thirty patients had bypass with a bubble oxygenator (Harvey H1700) and 34 patients had bypass with a flat sheet membrane oxygenator (COBE CML). In each case retinal fluorescein angiograms were obtained preoperatively and 5 minutes before the end of bypass and were processed with a digital image analyzer (Context Vision GOP-302). Microembolic perfusion defects were identified by digital subtraction of preoperative and end-bypass angiograms and their total area was computed. Results. In the bubble oxygenator group retinal perfusion defects indicative of microembolism occurred in all 30 (100%) patients. In contrast, over half the patients in the membrane oxygenator group had normal retinal perfusion, and the prevalence of perfusion defects (44%; 70% confidence limits 34% to 54%) was significantly less than in the bubble group (p less than 0.001). In addition, those patients in the membrane group had significantly fewer lesions (median 0; 70% confidence limits 0 to 1) than patients in the bubble group (median 2; 70% confidence limits 2 to 2; p less than 0.001) and also had significantly smaller total areas of retinal ischemia (median 0 mm2; 70% confidence limits 0 to 0.16 mm2) than the bubble group (median 0.22 mm2; 70% confidence limits 0.21 to 0.27 mm2; p less than 0.001). There was no relationship between the extent of retinal ischemia and bypass time, arterial blood gas concentrations, volume of cardiotomy suction or donor blood returned to the pump, or recent medication with aspirin. Inferences. Digital image analysis of retinal fluorescein angiograms may provide a method of quantifying microembolic ischemia in the central nervous system during cardiopulmonary bypass. Flat sheet membrane oxygenation appears to provide significantly better protection against microembolic ischemia than bubble oxygenation.     

Use of a membrane oxygenator for open-heart surgery in infants.

Because use of the bubble oxygenator during open-heart surgery is associated with complications such as hemolysis, pulmonary insufficiency and oliguria, a membrane oxygenator was used in conjunction with hypothermia in 37 infants. The main features of the oxygenator are gravitational blood flow, oxygenation into an airless, collapsible blood reservoir, low-flow roller pump flow back to the patient, accurate determination of flows and careful use of a heat exchanger. Gas flow (98% oxygen, 2% carbon dioxide) for the unit of 2 m2 is maintained at 3 to 4 1/min. Specific precautions are taken to ensure absence of bubbles. Three prime solutions are used, the final one having an osmolality of 381 mOsmol and containing 129.9 meq of sodium, 3.8 of potassium and 94.0 of chloride and 2001 mg/dl of glucose. Six patients died, but none of the deaths could be directly related to the use of the oxygenator. Respiratory complications were minimal, as were other complications. The technique is reliable in oxygenating blood in an tracorporeal circulation, but further familiarity with the membrane oxygenator for use in open-heart surgery in infants is desirable before firm conclusions can be drawn as to its value.     

Effect of oxygenator type and bypass flow pattern on the P(a-ET)CO2 gradient.

Changes in pre-bypass and post-bypass P(a-ET)CO2 gradients were evaluated regarding the type of bypass flow (pulsatile or nonpulsatile) and oxygenator (membrane or bubble). Duration of bypass and hemodynamic changes were analyzed also to determine their possible influence on PaCO2, PETCO2, and P(a-ET)CO2. A total of 36 adult patients undergoing cardiopulmonary bypass were anesthetized using a sufentanil-pancuronium-oxygen technique. Patients were divided into three groups based on the type of oxygenator and pump flow: group 1 (control group) consisted of a bubble oxygenator with nonpulsatile flow (BN), group 2 consisted of a bubble oxygenator with pulsatile flow (BP), and group 3 consisted of a membrane oxygenator with nonpulsatile flow (MN). Cardiac parameters (MAP, CI, SVR, and PVR) PaCO2, PETCO2, and P(a-ET)CO2 were determined pre-bypass and post-bypass following steady-state conditions. For the entire group there was a trend for the P(a-ET)CO2 gradient to increase in the post-bypass period (pre-bypass = 3.5 +/- 0.5 mm Hg, post-bypass = 4.3 +/- 0.5 mm Hg.). However, this increase was not statistically significant. Pulsatile flow (group 2) demonstrated a significant correlation with the change in P(a-ET)CO2 gradients from the pre-bypass to the post-bypass period (r = 0.85) when compared with the other two groups (group 1: r = -0.09 and group 3: r = 0.37). Thus, the P(a-ET)CO2 gradient tended to remain constant from the pre-bypass to the post-bypass period in group 2, whereas it increased in groups 1 and 3. Changes in MAP, CI, SVR, and PVR and the duration of CPB did not influence the P(a-ET)CO2 gradient.     

Avecor Trillium oxygenator versus noncoated Monolyth oxygenator: a prospective randomized controlled study

OBJECTIVES: The surface coating of a synthetic surface is currently investigated to decrease the harmful effects of cardiopulmonary bypass (CPB). This study was designed to study the effects of the surface coating of a hollow fiber membrane oxygenator on coagulation, inflammation markers, and clinical outcomes. The biomaterials used to coat the membrane include heparin, polyethylene oxide chains (PEO), and sulfate/sulfonate groups. The coated membrane was compared to an uncoated oxygenator made of polypropylene. METHODS: Two hundred patients who were scheduled to undergo valve repair and/or replacement surgery with or without coronary surgery were enrolled in the study. The patients were randomized to undergo CPB with either the Avecor oxygenator with Trillium (Medtronic, Minneapolis, MN, USA), a biopassive surface, or the Monolyth (Sorin, Irvine, CA, USA) oxygenator without coating. The primary and secondary endpoints were the differences between these oxygenators in regard to patients' biochemistry, coagulation profiles, inflammatory mediators, and clinical outcomes, including blood loss and neurological events. RESULTS: There were no differences between the two groups in terms of biochemistry, coagulation profile, inflammatory mediator release, and blood loss. Five patients in the Avecor group showed clinical evidence of a stroke confirmed with computerized tomography (CT) scan imaging, and none in the noncoated oxygenator group. CONCLUSION: The oxygenator Avecor offers similar results in terms of inflammation and coagulation profiles and blood loss during valvular surgery compared to a standard uncoated control oxygenator. The rate of neurological events was unusually elevated in the former group of patients, with only speculative explanation at this point. Further studies are warranted to clarify this aspect.     

Initial clinical experience with a low pressire drop membrane oxygenator for cardiopulmonary bypass in adult patients

The new Travenol oxygenator is composed of 80 parallel blood pathways. Microporous membrane separates the blood and gas compartments. The membrane surface area is 3 m², with a pore size of 0.01 microns. Venous blood drains directly from the patient through the oxygenator, then through an integral heat exchanger and into a reservoir, from which a single arterial pump returns the blood to the patient. The advantage of this configuration of membrane oxygenator is simplicity of setup and operation. A disadvantage that we have observed is an apparent variation in resistance to blood flow through the oxygenator during clinical perfusion. Construction changes in a later version of the oxygenator have reduced the resistance to flow through the blood pathway.This device has been used for 20 perfusions at moderate hypothermia (mean 31.8 °C) in patients up to 2.1 m² body surface area for up to 313 minutes. Blood flow was 2.1 to 5.6 liters/min, partial arterial oxygen pressure 100 to 394 torr, partial arterial carbon dioxide pressure 19 to 57 torr (mean 37 torr) and, arterial pH 7.29 to 7.56 (mean 7.41). Oxygen transfer was as high as 230 ml/min.This integral oxygenator-heat exchanger-reservoir is operated like a bubble oxygenator, with direct venous drainage through the device and a single pump, but it uses a membrane oxygenator for gas exchange to eliminate the detrimental effects of bubbles.     

Hematological advantage of a membrane oxygenator over a bubble oxygenator in long perfusions

To determine whether the large volumes of cardiotomy suction which occur during long perfusions can obscure the hematological advantage of the membrane oxygenator (MO) over the bubble oxygenator (BO), we studied 23 patients undergoing a coronary artery bypass grafting operation with an expected perfusion time of 3 hours (MO group, N = 10, SciMed spiral coil; BO group, N = 13, Shiley 100-A). During MO perfusion we found significantly higher platelet numbers, better platelet function (adenosine diphosphate-induced platelet aggregation), and less hemolysis (plasma hemoglobin), than during the BO perfusion. After the MO perfusion we measured significantly shorter bleeding times (Simplate II) and fewer transfusions of blood products. However, blood loss and whole-blood transfusions 18 hours after perfusion did not differ significantly between both groups. So in coronary artery bypass grafting operations with long perfusion times (mean, 3 hours), the MO still causes significantly less platelet and erythrocyte damage than the BO, despite the large volumes of cardiotomy suction known to occur during these operations.     

Reduced platelet activation and improved hemostasis after controlled cardiotomy suction during clinical membrane oxygenator perfusions

Platelet damage and postoperative blood loss are less severe after cardiopulmonary bypass performed with a membrane oxygenator than with a bubble oxygenator. However, this advantage of the membrane oxygenator can be partly negated by the platelet damage caused by cardiotomy suction, which implies the aspiration of air along with suction of blood. In order to reduce platelet damage by cardiotomy suction, we developed an automatic controlled cardiotomy suction system by which the aspiration of air was prevented. We evaluated platelet damage in a group of 28 patients (uncontrolled suction, n = 13; controlled suction, n = 15), and we studied the relationship between increasing volumes of cardiotomy suction and postoperative blood loss in a second group of 80 patients (uncontrolled suction, n = 47; controlled suction, n = 33). All patients underwent a coronary artery bypass operation with a membrane oxygenator. We measured significantly lower beta thromboglobulin concentrations during perfusions of approximately 2 hours and we observed a tendency toward shorter postoperative bleeding times if controlled cardiotomy suction was used. There were no significant differences between uncontrolled and controlled cardiotomy suction in platelet number and adenosine diphosphate-induced platelet aggregation. However, blood loss 18 hours postoperatively was significantly less in the controlled than in the uncontrolled suction group when the total volume of cardiotomy suction exceeded 65 L., which corresponded to perfusion times of over 3 hours. In conclusion, prevention of the aspiration of air along with suction of blood significantly reduced platelet activation and postoperative blood loss, particularly when large volumes of blood were aspirated.     

Deleterious effects of cardiopulmonary bypass. A prospective study of bubble versus membrane oxygenation

A number of hematologic and immunologic parameters that reflect erythrocyte and platelet damage and host defense mechanisms against infection were studied in 20 patients undergoing cardiopulmonary bypass during coronary operations. The patients were randomly assigned to a group in which a bubble oxygenator or a hollow-fiber membrane oxygenator was used. Hemolysis, thrombocytopenia, and significant release of beta thromboglobulin occurred in patients from the bubble oxygenator group and, to much lesser extent, in patients from the membrane oxygenator group. Polymorphonuclear leukocytes and monocytes from bubble oxygenator patients demonstrated increased generation of reactive oxygen species in the resting state and in the presence of the stimulating agents N-formyl-methionyl-leucyl-phenylalanine, concanavalin A, and opsonized zymosan, as compared with cells from membrane oxygenator patients. No difference was found between bubble and membrane oxygenator patients in the time of occurrence or intensity of leukopenia during bypass, of leukocytosis at the end of bypass, nor in the rate of complement activation, as assessed by quantitation of plasma C3a antigen. Complement activation was dependent on the alternative pathway. Immunoglobulin M concentration significantly decreased during bypass in both groups of patients. The serum opsonizing capacity for endotoxin and serum bactericidal activity for Serratia marcescens were decreased in both groups, mainly because of hemodilution, although they were additionally affected by bubble oxygenation. Several deleterious hematologic consequences of cardiopulmonary bypass can be minimized by the use of a membrane oxygenator. However, complement activation remains a potential risk factor even in membrane oxygenator patients and requires further investigation to obtain better hemocompatible materials for cardiopulmonary bypass circuits.     

Blood conservation with membrane oxygenators and dipyridamole

Cardiopulmonary bypass induces platelet activation and dysfunction, which result in platelet deposition and depletion. Reduced platelet numbers and abnormal platelet function may contribute to postoperative bleeding. A membrane oxygenator may preserve platelets and reduce bleeding more than a bubble oxygenator, and the antiplatelet agent dipyridamole may protect platelets intraoperatively and reduce bleeding postoperatively. A prospective randomized trial was performed in 44 patients undergoing elective coronary artery bypass grafting to assess the effects of the membrane oxygenator and dipyridamole on platelet counts, platelet activation products, and postoperative bleeding. Patients who were randomized to receive a bubble oxygenator and no dipyridamole had the lowest postoperative platelet counts, the greatest blood loss, and the most blood products transfused. Platelet counts were highest and blood loss was least in patients randomized to receive a membrane oxygenator and dipyridamole (p less than .05). A bubble oxygenator with dipyridamole and a membrane oxygenator without dipyridamole resulted in intermediate postoperative platelet counts and blood loss. Arterial thromboxane B2 and platelet factor 4 concentrations were elevated on cardiopulmonary bypass in all groups. Both the membrane oxygenator and dipyridamole were independently effective (by multivariate analysis) in preserving platelets. Optimal blood conservation was achieved with a membrane oxygenator and dipyridamole.     

Effects of prostaglandin E1 on platelet loss during in vivo and in vitro extracorporeal circulation with a bubble oxygenator.

Temporary inhibition of platelet function with prostaglandin E1 (PGE1) prevents platelets loss and functional alterations during extracorporeal circulation with a membrane oxygenator. This study evaluated the ability of PGE1 to prevent platelet injury in circuits containing a bubble oxygenator. During in vitro recirculation of human blood, the circulating platelet count, expressed as a percent of initial levels, decreased to 29%; platelets became insensitive to adenosine diphosphate (ADP) and to epinephrine; and plasma levels of low-affinity platelet factor 4 (LA-PF4) progressively rose to 7.4 microgram per milliliter. With PGE1 (1.2 micron), platelet counts remained stable at 92%; platelet reactivity remained normal for 1 hour; and plasma levels of LA-PF4 rose to only 3.3 microgram per milliliter. In rhesus monkeys that underwent cardiopulmonary bypass using a bubble oxygenator without PGE1, platelet counts, expressed as a percent of the prebypass platelet count, declined to 38%; platelets became insensitive to ADP; plasma levels of LA-PF4 progressively rose to 8.4 microgram per milliliter; and the mean postoperative bleeding time was 4.6 minutes. In monkeys that received PGE1, platelet counts declined to only 65%; platelets remained threefold more sensitive to ADP; platelets demonstrated delayed release of LA-PF4 and the mean postoperative bleeding time was 2.7 minutes. This report demonstrates that in a bubble oxygenator system, PGE1 reduces platelet loss, mitigates platelet injury, and shortens postoperative bleeding times following extracorporeal circulation.     

Blood cell trauma and postoperative bleeding: comparison of bubble and membrane oxygenators and observations on coronary suction

Blood cell trauma and postoperative bleeding were studied in 96 patients following coronary artery bypass grafting, with bubble oxygenator used in 47 cases and membrane oxygenator in 49. The haemocompatibility of membrane oxygenators was superior to that of the bubble type, as reflected by less haemolysis, better preservation of platelet function, less release of betathromboglobulin and less degranulation of neutrophil granulocytes. Coronary suction contributed to haemolysis, but did not affect platelet or granulocyte function. Fibrinolysis, postoperative blood loss and need for blood transfusion did not differ between the bubble and membrane oxygenator groups.     

A Comparison of the Effects of Membrane and Bubble Oxygenators on Platelet Counts and Platelet Size in Elective Cardiac Operations

To compare the effects of membrane and bubble oxygenators on platelet counts and the size of circulating platelets, serial hematocrits, platelet counts, and platelet sizing were measured in 23 patients undergoing elective cardiac operations. In 10 patients a bubble oxygenator was used and in 13, a SciMed membrane oxygenator. The two groups were statistically similar with respect to age, weight, time on bypass, and mean blood flow rates during bypass.It was found that platelet counts, when corrected for hemodilution, did not fall from control levels during or up to 24 hours after cardiopulmonary bypass in either group. In both groups, the relative number of platelets per gram of hemoglobin increased slightly during and after bypass, and this increase was significant in the bubble oxygenator group. The average size of circulating platelets increased only in the bubble oxygenator group, and then only in the one-day postoperative sample. These findings suggest that the membrane oxygenator offers no advantage with respect to preservation of platelets during cardiopulmonary bypass lasting up to 2 to 3 hours.