Oxygenation strategy and neurologic damage after deep hypothermic circulatory arrest. II. hypoxic versus free radical injury.

OBJECTIVES: Laboratory studies suggest that myocardial reperfusion injury is exacerbated by free radicals when pure oxygen is used during cardiopulmonary bypass. In phase I of this study we demonstrated that normoxic perfusion during cardiopulmonary bypass does not increase the risk of microembolic brain injury so long as a membrane oxygenator with an arterial filter is used. In phase II of this study we studied the hypothesis that normoxic perfusion increases the risk of hypoxic brain injury after deep hypothermia with circulatory arrest. METHODS: With membrane oxygenators with arterial filters, 10 piglets (8-10 kg) underwent 120 minutes of deep hypothermia and circulatory arrest at 15 degrees C, were rewarmed to 37 degrees C, and were weaned from bypass. In 5 piglets normoxia (PaO2 64-181 mm Hg) was used during cardiopulmonary bypass and in 5 hyperoxia (PaO2 400-900 mm Hg) was used. After 6 hours of reperfusion the brain was fixed for histologic evaluation. Near-infrared spectroscopy was used to monitor cerebral oxyhemoglobin and oxidized cytochrome a,a3 concentrations. RESULTS: Histologic examination revealed a significant increase in brain damage in the normoxia group (score 12.4 versus 8.6, P =.01), especially in the neocortex and hippocampal regions. Cytochrome a,a 3 and oxyhemoglobin concentrations tended to be lower during deep hypothermia and circulatory arrest in the normoxia group (P =.16). CONCLUSIONS: In the setting of prolonged deep hypothermia and circulatory arrest with membrane oxygenators, normoxic cardiopulmonary bypass significantly increases histologically graded brain damage with respect to hyperoxic cardiopulmonary bypass. Near-infrared spectroscopy suggests that the mechanism is hypoxic injury, which presumably overwhelms any injury caused by increased oxygen free radicals.     

Sequestration of glyceryl trinitrate (nitroglycerin) by cardiopulmonary bypass oxygenators

The effectiveness of glyceryl trinitrate (nitroglycerin) in controlling myocardial ischemia and blood pressure during coronary artery bypass graft surgery is frequently lost during surgery, possibly as a result of drug sequestration by the cardiopulmonary bypass circuit. The objective of this study was to utilize a gas-liquid chromatographic assay to determine the extent of removal of glyceryl trinitrate from the priming fluid by the bubble and membrane oxygenators. The apparatus was maintained at either 25 or 37 degrees C, the two extreme temperatures experienced by the patient during bypass surgery. At apparent steady state, the circulating glyceryl trinitrate concentration was decreased by 20.6%, 46.6%, and 67.3% with the Maxima membrane oxygenator, Cobe membrane oxygenator, and Bentley bubble oxygenator, respectively. The three-layer defoaming filters that are used in the Bentley bubble oxygenator were studied by immersing each of the three filters in fluid containing 60 nM glyceryl trinitrate and monitoring the drug concentration in Plasmalyte. The filters sequestered approximately 90% of the glyceryl trinitrate from the bathing solution of which 31% was recovered with a single methanol wash of the polyurethane filter. These data demonstrate that the different oxygenators used in the cardiopulmonary bypass circuit remove glyceryl trinitrate to varying degrees from the circulating fluid.     

Complement activation during cardiopulmonary bypass. Comparison of bubble and membrane oxygenators

A prospective randomized trial involving 91 patients undergoing cardiopulmonary bypass compared the effects of bubble oxygenators (with and without methylprednisolone sodium succinate) and membrane oxygenators on complement activation and transpulmonary sequestration of leukocytes. Patients were divided as follows: Group I, 30 patients, bubble oxygenator; Group II, 31 patients, bubble oxygenator and methylprednisolone sodium succinate (30 mg/kg); Group III, 30 patients, membrane oxygenator. In Group I, C3a increased from 323 +/- 171 ng/ml during cardiopulmonary bypass to 1,564 +/- 785 ng/ml at 25 minutes after bypass (p less than 0.0001). A significant decrease in C3a was found in Groups II and III compared to Group I (p less than 0.0001). C5a did not change significantly during cardiopulmonary bypass in any group. Reestablishment of pulmonary circulation at the end of bypass produced significant transpulmonary leukocyte sequestration in Group I; the median cell difference was 1,700/microliter. Transpulmonary sequestration was significantly (p less than 0.0001) less in Group II (median cell difference = 200/microliter) and in Group III (median cell difference = 400/microliter) than in Group I. We conclude that cardiopulmonary bypass with a bubble oxygenator alone initiates significantly (p less than 0.0001) more C3a activation and leukocyte sequestration than when methylprednisolone sodium succinate (30 mg/kg) is given 20 minutes before the start of cardiopulmonary bypass with a bubble oxygenator or when a silicone membrane oxygenator is used.     

The effect of arterial filtration on reduction of gaseous microemboli in the middle cerebral artery during cardiopulmonary bypass

Noninvasive in vivo detection of gaseous microemboli in the middle cerebral artery, by transcranial Doppler ultrasound, was used to determine the effect of filtration in the arterial catheter using 25- and 40-microns filters and bubble oxygenators in patients undergoing cardiopulmonary bypass surgery. Eighteen patients undergoing coronary artery bypass surgery were studied using a closed cardiac (unvented heart) model. Group 1 patients (no filters) had the highest incidence of gaseous microemboli, as indicated by the ultrasound microemboli index, at both high and low oxygen flow rates. Group 2 patients (40-microns filters) had a significantly lower microemboli index, particularly at low oxygen flow rates (t = 4.9, p less than 0.001). The 25-microns group patients had the lowest values of all. No microemboli were detected at low oxygen flow rates, and microemboli were detected in only 0.1% of the samples at high oxygen flow rates. Additionally, observations on vented hearts in 3 patients undergoing cardiac valve surgery indicate that the origin of gaseous microemboli may be air trapped inside the heart.     

Impact of oxygenator characteristics on its capability to remove gaseous microemboli

Since the advent of cardiopulmonary bypass, the generation and elimination of gaseous and solid (micro) emboli have been a concern. Major improvements with respect to gaseous microemboli have been made by the introduction of arterial line filtration and membrane oxygenators. Animal experiments have shown a clear correlation between massive air embolism and outcome. However, limited knowledge is available regarding the cut-off point between the occurrence of negative outcome and the number and size of gaseous microemboli. Generation of gaseous microemboli can occur when using cardiopulmonary bypass. However, no consensus exists on when a given diameter or number of emboli becomes injurious to the patient. An important variable is the gas mixture inside the bubble. Nitrogen has a very long dissolution time that results in a prolonged ischemia for tissue behind the occlusion. The pathophysiologic reaction of the body when exposed to gaseous microemboli is most likely based on ischemia caused by partial occlusion of blood vessels and by endothelial damage. Gaseous microemboli can be cleared mechanically by using filters, by reduction of blood velocity, and by rapid reduction of the nitrogen content. Elimination of gaseous microemboli is dependent on the design of the cardiopulmonary bypass circuit. A membrane oxygenator, although not designed for it, can remove gaseous microemboli. Arterial line filtration is not the best solution for removal of gaseous microemboli, because larger emboli have been fractionated before reaching the arterial filter. Venous line filtration is a more efficient way for clearing gaseous microemboli.     

Comparison of gaseous microemboli counts in arterial, simultaneous and venous heat exchange with a hollow fiber membrane oxygenator

Potential sources of gaseous microemboli during cardiopulmonary bypass are varied. However, it is known that membrane oxygenators generate fewer gaseous microemboli than bubble oxygenators and that bubblers cannot utilize arterial heat exchange without generating significant gaseous microemboli during rewarming. A membrane oxygenator utilizing simultaneous gas and heat exchange raises the concern that concurrent gas and heat exchange would result in a higher production of gaseous microemboli compared to conventional venous heat exchange devices. This in vitro study compared venous, simultaneous, arterial and control (venous) heat exchanger gaseous microemboli counts during rewarming. No significant difference was found between the four heat exchangers when comparing inlet and outlet gaseous microemboli counts. This in vitro study suggests that there is no difference in gaseous microemboli generation when varying the position of the heat exchanger in the extracorporeal circuit incorporating a microporous membrane oxygenator.     

Comparative evaluation of a new disposable rotating membrane oxygenator with bubble oxygenator.

The comparative in vivo performance of adult-size bubble and rotating membrane oxygenators was evaluated during closed-chest cardiopulmonary bypass for six hours in two groups of dogs. The results show that the rotating membrane oxygenator is efficient in oxygen and carbon dioxide transfer with minimal trauma to blood, while platelet destruction and hemolysis were marked with the bubble oxygenator. Cerebral, cardiac, and respiratory complications were frequent with the bubble oxygenator and absent with the membrane oxygenator.     

The detection of microemboli in the middle cerebral artery during cardiopulmonary bypass: a transcranial Doppler ultrasound investigation using membrane and bubble oxygenators

Twenty-seven patients were examined who were undergoing cardiopulmonary bypass (CPB) surgery with either a bubble oxygenator or a capillary membrane oxygenator. The latter incorporated an arterial filter and bubble trap. A noninvasive Doppler ultrasound technique is described for monitoring irregularities in the Doppler flow signals attributable to gaseous microemboli detected in the middle cerebral artery during CPB. The ultrasound index for detecting gaseous microemboli (MEI) indicated the presence of such microemboli in 22 of the 27 patients during insertion of the aortic cannula. Measurements during CPB showed the MEI ranged from 4 to 39 in the 17 patients with a bubble oxygenator. However, all 10 patients with a membrane oxygenator had an MEI of 0. Varying the gas flow rates in 3 patients with bubble oxygenators showed a change in MEI from 4 +/- 4 (SD) at a flow rate of 2 L/min to 17 +/- 9 at a flow rate of 5 L/min. This observation supports the assumption that the MEI is providing quantitative information regarding the presence of gaseous emboli in the middle cerebral artery.     

Ultrasonic identification of sources of gaseous microemboli during open heart surgery

An ultrasonic transducer was used to detect gaseous microemboli during open heart surgery. When placed over the carotid artery of patients during surgery it was recognized that gaseous emboli were released into the systemic circulation following recognizable surgical procedures and from the oxygenator during bypass. Fixation of the transducer on the arterial return from the oxygenator demonstrated that the two types of disposable bubble oxygenators and disc oxygenators tested delivered microbubbles in the arterial blood. Some factors affecting the number of bubbles released were recognized and are discussed. Case reports of two patients who received large volumes of gaseous microemboli from different sources are presented.     

Different oxygenators for cardiopulmonary bypass lead to varying degrees of human complement activation in vitro

Complement activation was studied in vitro with six different membrane and bubble oxygenators for cardiopulmonary bypass. There was a similar increase in terminal (C5 to C9) activation with all oxygenators (p less than 0.001), ranging from 281% (117% to 444%) to 453% (225% to 680%) after 60 minutes (median and 95% confidence intervals). C3 activation was not observed with a hollow fiber membrane and a soft shell bubble oxygenator. On the other hand, a capillary membrane, a sheet membrane, a nonporous membrane, and a hard shell bubble oxygenator all induced a similar increase in C3 activation (p less than 0.01), ranging from 107% (23% to 346%) to 272% (88% to 395%) after 60 minutes. The differences in C3 activation could not be explained by the blood contact materials or any other single factor known to induce activation, which suggests that overall complement activation during cardiopulmonary bypass is a multifactorial effect. The tubing set per se induced only minor C3 activation but contributed to the overall formation of terminal complement complex. The study further indicates that an arterial line blood filter prevents activated neutrophils from being reinfused to the patient and should be used regardless of type of oxygenator.     

Quantitation of particulate microemboli during cardiopulmonary bypass: experimental and clinical studies.

An electronic particle-size analyzer (Coulter Counter ZM) was used to quantitate particulate microemboli 15 to 80 microns in size during cardiopulmonary bypass. Through both laboratory studies and clinical research, we confirmed three main causes of microemboli: (1) infusion of banked blood stored for more than 3 days; (2) use of cardiotomy reservoirs; and (3) use of bubble oxygenators. The regression equation between number of particles and blood storage time was Y = 3.7262X + 10.244 (r = 0.886; p < 0.01). The number of microemboli from cardiotomy reservoirs was 2.8 to 5.1 times that from other sources (p < 0.01). The number of solid particles from bubble oxygenators was 1.8 to 3.2 times that from membrane oxygenators (p < 0.01). Electron microscopy showed that a large number of solid particles more than 20 microns in size were formed during heart-lung bypass. They obstructed the microcirculation and damaged pulmonary capillary endothelial and alveolar epithelial cells. The degree of histological damage was related to the number and size of microemboli and the duration of pulmonary microcirculatory obstruction.     

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.     

Similarity of Clinical and Laboratory Results Obtained with Microporous Teflon Membrane Oxygenator and Bubble-Film Hybrid Oxygenator

For 80 elective clinical cardiopulmonary bypasses we alternately used either a commercial microporous Teflon membrane oxygenator or a commercial hybrid bubble-film oxygenator. Setup time was a little longer with the membrane unit (20 minutes), but priming volume (2,250 ml) was the same. No problems were encountered with the hybrid oxygenator. However, despite our monitoring of additional variables, including shim and inlet pressure and recirculation flow, gas exchange abnormalities were encountered in 5 patients on whom the membrane oxygenator was used; in 4 of these cases the abnormalities were encountered prior to our recognition of the potential for occasional internal shunting with this device.There were no hospital deaths. When the two groups, matched except for oxygenator selection, were compared, there were no significant differences clinically or hematologically. For cardiopulmonary bypass of 2 hours or less, both oxygenators studied are definite improvements over previous silicone membrane and high-gas-flow bubble oxygenators. However, lower cost and reduced complexity favor the hybrid oxygenator.     

Significance of the concentrated red cell and albumin priming method with particular reference to anaphylatoxin generation.

Anaphylatoxins generated by cardiopulmonary bypass were observed in basic and clinical studies (n = 120 in the latter). In vitro immunoglobulin fractions denatured by oxygen bubbling produced C4a, C3a, and C5a, but albumin identically treated did not. Therefore concentrated red cells with albumin were used to prime homologous blood for clinical application during cardiopulmonary bypass. Complement levels were compared with type of oxygenator (bubble or membrane) and the ratio of primed homologous blood to circulating autologous blood volume. With the bubble oxygenator at a low ratio of homologous to autologous blood (arbitrarily defined as less than 20%), C3a levels during cardiopulmonary bypass tended to be lower in the concentrated red cells plus albumin priming group than in the ordinary priming group (p less than 0.1, at 60 and 90 minutes of cardiopulmonary bypass). C4a and C3a levels increased less after protamine administration with concentrated red cells plus albumin priming (p less than 0.05, p less than 0.01, respectively, 90 minutes after protamine) than with ordinary priming. Such changes in the membrane oxygenator group were less remarkable. Thus C3a levels were approximately the same in both oxygenator groups primed with concentrated red cells plus albumin. The higher the homologous to autologous ratio, the steeper the C4a and C3a increase from the beginning of cardiopulmonary bypass with the bubble oxygenator. This tendency was less remarkable in the membrane oxygenator group. Early postoperative pulmonary function was improved by concentrated red cells plus albumin priming, especially in the bubble oxygenator group. In conclusion, (1) oxygenator systems primed with concentrated red cells plus albumin produced less anaphylatoxin than those with homologous blood, especially with the bubble oxygenator, and (2) our clinical results support the importance of immunoglobulin denatured by oxygen bubbling in anaphylatoxin generation (by means of the classical pathway), as shown by our in vitro study.     

Hypothermia: its possible role in cardiac surgery

The current safety of operations on the heart requiring cardiopulmonary bypass occurred because of a series of step-by-step laboratory and clinical investigations that were compromises between the time needed for heart repair and the brain's requirement for oxygen. The first step, so clearly shown in a paper by Bigelow and associates in 1950, was the reduction of the brain's need for oxygen by surface cooling to 28 degrees to 32 degrees C, limited to this level by cardiac and pulmonary failure at levels lower than this. The six to eight minutes of circulatory arrest permitted time for repair of simple defects. This method was rapidly adopted by many surgeons. As low-flow pump oxygenators became available, blood cooling to 10 degrees to 20 degrees C was introduced. This increased the periods of circulatory arrest to 30 to 60 minutes, and also made still longer periods of bypass with the pump oxygenator possible. Hypothermia to reduce oxygen and metabolic requirements is still an important adjunct to bypass, even with the currently used efficient pump oxygenators. It remains the most important component of myocardial preservation, and has made possible the delay needed for transportation between the harvesting and the transplantation of organs.     

Comparison of bubble and membrane oxygenators in short and long perfusions.

Eighty patients had cardiopulmonary bypass (CPB), half having short (109 +/- 11 minutes) perfusions and half having long (188 +/- 14 min) perfusions. Twenty patients in each group were perfused with bubble oxygenators (Bentley, Harvey, or Galen) and 20 with membrane oxygenators (Modulung or Teflo). Hemodilution to a hematocrit value of 22.5% +/- 1.4% and hypothermia to 28 degrees +/- 2 degrees C were used in all patients. Complete hemograms, sequential multiple analyzer 18 tests, coagulation profiles, blood gases and pH, three immunoglobulins, and two complement fraction proteins were sampled as follows: three times before perfusion, one to ten times during perfusion, 1 hour immediately after perfusion, and 4, 24, and 48 hours postoperatively. Data in concentration terms were compared statistically and reported as mean and standard error for each subset. Additionally, rates of gain or loss were calculated in terms of quantity per liter of blood pumped per minute. During perfusion for both duration sets, use of a membrane oxygenator resulted in greater pump flows (4.55 +/- 0.15 L/min versus 3.75 +/- 0.11 L/min), lower total peripheral resistances (1,125 +/- 63 dynes.sec.cm-5 versus 1,652 +/- 115 dynes.sec.cm-5), and greater urinary outputs (9.4 +/- 1.1 ml/min versus 2.2 +/- 0.6 ml/min) than in the bubble oxygenator subsets. Comparisons of measured and calculated data in the immediate postperfusion interval showed no differences between bubble and membrane oxygenator subsets for short perfusions. In long perfusions, the membrane subset had lower plasma hemoglobin and white cell concentrations and generation rates, smaller (3 to 8 1/2 times) losses of IgG, IgM, C3 and shed blood necessitating less transfusion, and greater C4 losses. The membrane oxygenator systems used were more complex and costly and offered no advantages for short perfusion in adults. In anticipated long perfusions or where bleeding may be a problem, a membrane oxygenator appears more efficacious than bubble systems. For perfusions of less than 2 hours, membrane oxygenators had no biochemical or hematologic advantage over the bubble devices used in this study.     

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.     

Insulin-like growth factor 1 improves the relationship between systemic oxygen consumption and delivery in piglets after cardiopulmonary bypass

OBJECTIVE: We sought to assess the effects of insulin-like growth factor 1 on the balance between systemic oxygen consumption and oxygen delivery after cardiopulmonary bypass in piglets. METHODS: Twelve piglets weighing 4.5 to 8.3 kg undergoing hypothermic (28 degrees C) cardiopulmonary bypass for 70 to 120 minutes with 40 minutes of aortic crossclamping were studied before and during the first 6 hours after cardiopulmonary bypass. Oxygen consumption was continuously measured by an indirect calorimeter, Deltatrac II MBM-200 Metabolic Monitor (Datex Division Instrumentarium, Helsinki, Finland). Oxygen delivery and cardiac output were calculated from oxygen consumption and the arterial and mixed venous oxygen contents sampled before and every 30 minutes after cardiopulmonary bypass. Oxygen extraction ratio was derived by the ratio of oxygen consumption to oxygen delivery. Arterial blood lactate was measured before and every 30 minutes after cardiopulmonary bypass. Six animals were randomly assigned to receive an intravenous infusion of insulinlike growth factor 1 at 1.2 mg/h from 1 to 6 hours after cardiopulmonary bypass; the remaining 6 served as a control group. RESULTS: Relative to the control group, intravenous infusion of insulin-like growth factor 1 significantly reduced oxygen consumption (P =.02) and increased cardiac output (P =.016) and oxygen delivery (P =.049) during the first 6 hours after surgery with hypothermic cardiopulmonary bypass. As a result, oxygen extraction was significantly decreased (P =.012). CONCLUSIONS: Intravenous infusion of insulin-like growth factor 1 improved oxygen transport by reducing oxygen consumption as well as increasing cardiac output and oxygen delivery during the first 6 hours after cardiopulmonary bypass in piglets. This may have important clinical implications for the care of critically ill children after surgery with cardiopulmonary bypass.     

Complement activation with bubble and membrane oxygenators in aortocoronary bypass grafting

Thirty-three patients admitted for coronary bypass grafting were randomized to cardiopulmonary bypass with a bubble oxygenator (Cobe or Polystan) or a membrane oxygenator (SciMed). Plasma concentrations of C3 activation products and the terminal complement complex were measured using enzyme immunoassays. Both variables increased almost linearly after onset of cardiopulmonary bypass, with maximal concentrations at closure of the sternum. From a baseline of 7.5 to 12.0 arbitrary units (AU)/mL (medians), the concentrations of C3 activation products increased by 117.5 AU/mL (Cobe), 120.5 AU/mL (Polystan), and 213.3 AU/mL (SciMed). The increase in the membrane group was significantly higher than in the two bubble oxygenator groups (p less than 0.01). From a baseline of 0.9 to 1.3 AU/mL, the concentrations of terminal complement complex increased by 5.4 AU/mL (Cobe), 6.6 AU/mL (Polystan), and 7.7 AU/mL (SciMed) (differences not significant). The higher C3 activation caused by the membrane oxygenator may be explained by differences in flow profile and surface area in contact with blood. The study cannot confirm the general assumption that membrane oxygenators lead to lower complement activation than do bubble oxygenators.     

Pressure drop, shear stress, and activation of leukocytes during cardiopulmonary bypass: a comparison between hollow fiber and flat sheet membrane oxygenators

The membrane oxygenator is known to be superior to the bubble oxygenator, but little information is available about the difference between the hollow fiber and flat sheet membrane oxygenators with regard to pressure drop, shear stress, and leukocyte activation. In this study, we compared these 2 types of membrane oxygenators in patients undergoing cardiopulmonary bypass (CPB) surgery with special focus on leukocyte activation and pressure drop across the oxygenators. Plasma concentration of elastase, a marker indicating leukocyte activation, increased to 593+/-68% in the flat sheet oxygenator group versus 197+/-42% in the hollow fiber oxygenator group (p<0.01) at the end of CPB compared to their respective baseline concentrations before CPB. Pressure drop across the oxygenator was significantly higher in the flat sheet group than in the hollow fiber group throughout the entire period of CPB (p<0.01). High pressure drop across the oxygenator as well as the calculated shear stress was positively correlated with the release of elastase at the end of CPB (r = 0.760, p<0.01, r = 0.692, p<0.01). However, this positive correlation existed in the flat sheet oxygenator but not in the hollow fiber oxygenator. Clinically, both membrane oxygenators have satisfactory performance in O2 and CO2 transfer. These results suggest that a higher pressure drop across the flat sheet oxygenator is associated with more pronounced activation of leukocytes in patients undergoing cardiopulmonary bypass.